Prepare 0.9.1 release (#3448)

Co-authored-by: Francesco Renzi <rentziass@gmail.com>

.github/ISSUE_TEMPLATE/config.yml

@@ -1,5 +1,8 @@
 blank_issues_enabled: false
 contact_links:
+- name: Feature requests for the gha-runner-scale-set (actions.github.com API group)
+  about: Feature requests associated with the actions.github.com group should be posted on the GitHub Community Support Forum
+  url: https://github.com/orgs/community/discussions/categories/actions
 - name: Sponsor ARC Maintainers
   about: If your business relies on the continued maintainance of actions-runner-controller, please consider sponsoring the project and the maintainers.
   url: https://github.com/actions/actions-runner-controller/tree/master/CODEOWNERS

.github/ISSUE_TEMPLATE/github_bug_report.yaml

@@ -0,0 +1,113 @@
+name: Bug Report (actions.github.com API group)
+description: File a bug report for actions.github.com API group
+title: "<Please write what didn't work for you here>"
+labels: ["bug", "needs triage", "gha-runner-scale-set"]
+body:
+- type: checkboxes
+  id: read-troubleshooting-guide
+  attributes:
+    label: Checks
+    description: Please check all the boxes below before submitting
+    options:
+    - label: I've already read https://docs.github.com/en/actions/hosting-your-own-runners/managing-self-hosted-runners-with-actions-runner-controller/troubleshooting-actions-runner-controller-errors and I'm sure my issue is not covered in the troubleshooting guide.
+      required: true
+
+    - label: I am using charts that are officially provided
+- type: input
+  id: controller-version
+  attributes:
+    label: Controller Version
+    description: Refers to semver-like release tags for controller versions. Any release tags prefixed with `gha-runner-scale-set-` are releases associated with this API group
+    placeholder: ex. 0.6.1
+  validations:
+    required: true
+- type: dropdown
+  id: deployment-method
+  attributes:
+    label: Deployment Method
+    description: Which deployment method did you use to install ARC?
+    options:
+      - Helm
+      - Kustomize
+      - ArgoCD
+      - Other
+  validations:
+    required: true
+- type: checkboxes
+  id: checks
+  attributes:
+    label: Checks
+    description: Please check all the boxes below before submitting
+    options:
+    - label: This isn't a question or user support case (For Q&A and community support, go to [Discussions](https://github.com/actions/actions-runner-controller/discussions)).
+      required: true
+    - label: I've read the [Changelog](https://github.com/actions/actions-runner-controller/blob/master/docs/gha-runner-scale-set-controller/README.md#changelog) before submitting this issue and I'm sure it's not due to any recently-introduced backward-incompatible changes
+      required: true
+- type: textarea
+  id: reproduction-steps
+  attributes:
+    label: To Reproduce
+    description: "Steps to reproduce the behavior"
+    render: markdown
+    placeholder: |
+      1. Go to '...'
+      2. Click on '....'
+      3. Scroll down to '....'
+      4. See error
+  validations:
+    required: true
+- type: textarea
+  id: actual-behavior
+  attributes:
+    label: Describe the bug
+    description: Also tell us, what did happen?
+    placeholder: A clear and concise description of what happened.
+  validations:
+    required: true
+
+- type: textarea
+  id: expected-behavior
+  attributes:
+    label: Describe the expected behavior
+    description: Also tell us, what did you expect to happen?
+    placeholder: A clear and concise description of what the expected behavior is.
+  validations:
+    required: true
+
+- type: textarea
+  id: additional-context
+  attributes:
+    label: Additional Context
+    render: yaml
+    description: |
+      Provide `values.yaml` files that are relevant for this issue. PLEASE REDACT ANY INFORMATION THAT SHOULD NOT BE PUBLICALY AVAILABLE, LIKE GITHUB TOKEN FOR EXAMPLE.
+    placeholder: |
+      PLEASE REDACT ANY INFORMATION THAT SHOULD NOT BE PUBLICALY AVAILABLE, LIKE GITHUB TOKEN FOR EXAMPLE.
+  validations:
+    required: true
+
+- type: textarea
+  id: controller-logs
+  attributes:
+    label: Controller Logs
+    description: "NEVER EVER OMIT THIS! Include complete logs from `actions-runner-controller`'s controller-manager pod."
+    render: shell
+    placeholder: |
+      PROVIDE THE LOGS VIA A GIST LINK (https://gist.github.com/), NOT DIRECTLY IN THIS TEXT AREA
+
+      To grab controller logs:
+
+      kubectl logs -n $NAMESPACE deployments/$CONTROLLER_DEPLOYMENT
+  validations:
+    required: true
+- type: textarea
+  id: runner-pod-logs
+  attributes:
+    label: Runner Pod Logs
+    description: "Include logs and kubectl describe output from runner pod(s)."
+    render: shell
+    placeholder: |
+      PROVIDE THE WHOLE LOGS VIA A GIST LINK (https://gist.github.com/), NOT DIRECTLY IN THIS TEXT AREA
+  validations:
+    required: true
+

.github/ISSUE_TEMPLATE/summerwind_bug_report.yaml

@@ -1,7 +1,7 @@
-name: Bug Report
-description: File a bug report
+name: Bug Report (actions.summerwind.net API group)
+description: File a bug report for actions.summerwind.net API group
 title: "<Please write what didn't work for you here>"
-labels: ["bug", "needs triage"]
+labels: ["bug", "needs triage", "community"]
 body:
 - type: checkboxes
   id: read-troubleshooting-guide

.github/ISSUE_TEMPLATE/summerwind_feature_request.md

@@ -1,7 +1,7 @@
 ---
-name: Feature request
+name: Feature request (actions.summerwind.net API group)
 about: Suggest an idea for this project
-labels: ["enhancement", "needs triage"]
+labels: ["enhancement", "needs triage", "community"]
 title: ''
 assignees: ''
 ---

.github/actions/execute-assert-arc-e2e/action.yaml

@@ -193,7 +193,7 @@ runs:
       shell: bash
       run: |
         helm uninstall ${{ inputs.arc-name }} --namespace ${{inputs.arc-namespace}} --debug
-        kubectl wait --timeout=10s --for=delete AutoScalingRunnerSet -n ${{inputs.arc-name}} -l app.kubernetes.io/instance=${{ inputs.arc-name }}
+        kubectl wait --timeout=30s --for=delete AutoScalingRunnerSet -n ${{inputs.arc-namespace}} -l app.kubernetes.io/instance=${{ inputs.arc-name }}
 
     - name: Gather logs and cleanup
       shell: bash

.github/workflows/arc-publish-chart.yaml

@@ -63,7 +63,7 @@ jobs:
         python-version: '3.11'
 
     - name: Set up chart-testing
-      uses: helm/chart-testing-action@v2.3.1
+      uses: helm/chart-testing-action@v2.6.0
 
     - name: Run chart-testing (list-changed)
       id: list-changed

.github/workflows/arc-release-runners.yaml

@@ -17,7 +17,7 @@ env:
   PUSH_TO_REGISTRIES: true
   TARGET_ORG: actions-runner-controller
   TARGET_WORKFLOW: release-runners.yaml
-  DOCKER_VERSION: 20.10.23
+  DOCKER_VERSION: 24.0.7
 
 concurrency:
   group: ${{ github.workflow }}

.github/workflows/arc-update-runners-scheduled.yaml

@@ -126,15 +126,19 @@ jobs:
 
       - name: Update files
         run: |
-          sed -i "s/$RUNNER_CURRENT_VERSION/$RUNNER_LATEST_VERSION/g" runner/VERSION
-          sed -i "s/$RUNNER_CURRENT_VERSION/$RUNNER_LATEST_VERSION/g" runner/Makefile
-          sed -i "s/$RUNNER_CURRENT_VERSION/$RUNNER_LATEST_VERSION/g" Makefile
-          sed -i "s/$RUNNER_CURRENT_VERSION/$RUNNER_LATEST_VERSION/g" test/e2e/e2e_test.go
+          CURRENT_VERSION="${RUNNER_CURRENT_VERSION//./\\.}"
+          LATEST_VERSION="${RUNNER_LATEST_VERSION//./\\.}"
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" runner/VERSION
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" runner/Makefile
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" Makefile
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" test/e2e/e2e_test.go
 
-          sed -i "s/$CONTAINER_HOOKS_CURRENT_VERSION/$CONTAINER_HOOKS_LATEST_VERSION/g" runner/VERSION
-          sed -i "s/$CONTAINER_HOOKS_CURRENT_VERSION/$CONTAINER_HOOKS_LATEST_VERSION/g" runner/Makefile
-          sed -i "s/$CONTAINER_HOOKS_CURRENT_VERSION/$CONTAINER_HOOKS_LATEST_VERSION/g" Makefile
-          sed -i "s/$CONTAINER_HOOKS_CURRENT_VERSION/$CONTAINER_HOOKS_LATEST_VERSION/g" test/e2e/e2e_test.go
+          CURRENT_VERSION="${CONTAINER_HOOKS_CURRENT_VERSION//./\\.}"
+          LATEST_VERSION="${CONTAINER_HOOKS_LATEST_VERSION//./\\.}"
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" runner/VERSION
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" runner/Makefile
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" Makefile
+          sed -i "s/$CURRENT_VERSION/$LATEST_VERSION/g" test/e2e/e2e_test.go
 
       - name: Commit changes
         run: |

.github/workflows/arc-validate-chart.yaml

@@ -69,10 +69,10 @@ jobs:
       # python is a requirement for the chart-testing action below (supports yamllint among other tests)
       - uses: actions/setup-python@v4
         with:
-          python-version: '3.7'
+          python-version: '3.11'
 
       - name: Set up chart-testing
-        uses: helm/chart-testing-action@v2.4.0
+        uses: helm/chart-testing-action@v2.6.0
 
       - name: Run chart-testing (list-changed)
         id: list-changed

.github/workflows/gha-e2e-tests.yaml

@@ -16,7 +16,7 @@ env:
   TARGET_ORG: actions-runner-controller
   TARGET_REPO: arc_e2e_test_dummy
   IMAGE_NAME: "arc-test-image"
-  IMAGE_VERSION: "0.6.1"
+  IMAGE_VERSION: "0.9.1"
 
 concurrency:
   # This will make sure we only apply the concurrency limits on pull requests
@@ -880,3 +880,98 @@ jobs:
           helm uninstall "${{ steps.install_arc.outputs.ARC_NAME }}" --namespace "arc-runners" --debug
           kubectl wait --timeout=10s --for=delete AutoScalingRunnerSet -n "${{ steps.install_arc.outputs.ARC_NAME }}" -l app.kubernetes.io/instance="${{ steps.install_arc.outputs.ARC_NAME }}"
           kubectl logs deployment/arc-gha-rs-controller -n "arc-systems"
+
+  init-with-min-runners:
+    runs-on: ubuntu-latest
+    timeout-minutes: 20
+    if: github.event_name != 'pull_request' || github.event.pull_request.head.repo.id == github.repository_id
+    env:
+      WORKFLOW_FILE: arc-test-workflow.yaml
+    steps:
+      - uses: actions/checkout@v3
+        with:
+          ref: ${{ github.head_ref }}
+
+      - uses: ./.github/actions/setup-arc-e2e
+        id: setup
+        with:
+          app-id: ${{secrets.E2E_TESTS_ACCESS_APP_ID}}
+          app-pk: ${{secrets.E2E_TESTS_ACCESS_PK}}
+          image-name: ${{env.IMAGE_NAME}}
+          image-tag: ${{env.IMAGE_VERSION}}
+          target-org: ${{env.TARGET_ORG}}
+
+      - name: Install gha-runner-scale-set-controller
+        id: install_arc_controller
+        run: |
+          helm install arc \
+          --namespace "arc-systems" \
+          --create-namespace \
+          --set image.repository=${{ env.IMAGE_NAME }} \
+          --set image.tag=${{ env.IMAGE_VERSION }} \
+          --set flags.updateStrategy="eventual" \
+          ./charts/gha-runner-scale-set-controller \
+          --debug
+          count=0
+          while true; do
+            POD_NAME=$(kubectl get pods -n arc-systems -l app.kubernetes.io/name=gha-rs-controller -o name)
+            if [ -n "$POD_NAME" ]; then
+              echo "Pod found: $POD_NAME"
+              break
+            fi
+            if [ "$count" -ge 60 ]; then
+              echo "Timeout waiting for controller pod with label app.kubernetes.io/name=gha-rs-controller"
+              exit 1
+            fi
+            sleep 1
+            count=$((count+1))
+          done
+          kubectl wait --timeout=30s --for=condition=ready pod -n arc-systems -l app.kubernetes.io/name=gha-rs-controller
+          kubectl get pod -n arc-systems
+          kubectl describe deployment arc-gha-rs-controller -n arc-systems
+
+      - name: Install gha-runner-scale-set
+        id: install_arc
+        run: |
+          ARC_NAME=${{github.job}}-$(date +'%M%S')$((($RANDOM + 100) % 100 + 1))
+          helm install "$ARC_NAME" \
+          --namespace "arc-runners" \
+          --create-namespace \
+          --set githubConfigUrl="https://github.com/${{ env.TARGET_ORG }}/${{env.TARGET_REPO}}" \
+          --set githubConfigSecret.github_token="${{ steps.setup.outputs.token }}" \
+          --set minRunners=5 \
+          ./charts/gha-runner-scale-set \
+          --debug
+          echo "ARC_NAME=$ARC_NAME" >> $GITHUB_OUTPUT
+          count=0
+          while true; do
+            POD_NAME=$(kubectl get pods -n arc-systems -l actions.github.com/scale-set-name=$ARC_NAME -o name)
+            if [ -n "$POD_NAME" ]; then
+              echo "Pod found: $POD_NAME"
+              break
+            fi
+            if [ "$count" -ge 60 ]; then
+              echo "Timeout waiting for listener pod with label actions.github.com/scale-set-name=$ARC_NAME"
+              exit 1
+            fi
+            sleep 1
+            count=$((count+1))
+          done
+          kubectl wait --timeout=30s --for=condition=ready pod -n arc-systems -l actions.github.com/scale-set-name=$ARC_NAME
+          kubectl get pod -n arc-systems
+      - name: Ensure 5 runners are up
+        run: |
+          count=0
+          while true; do
+            pod_count=$(kubectl get pods -n arc-runners --no-headers | wc -l)
+            if [[ "$pod_count" = 5 ]]; then
+              echo "5 pods are up!"
+              break
+            fi
+            if [[ "$count" -ge 12 ]]; then
+              echo "Timeout waiting for 5 pods to be created"
+              exit 1
+            fi
+            sleep 1
+            count=$((count+1))
+          done

.github/workflows/gha-validate-chart.yaml

@@ -65,10 +65,10 @@ jobs:
       # python is a requirement for the chart-testing action below (supports yamllint among other tests)
       - uses: actions/setup-python@v4
         with:
-          python-version: '3.7'
+          python-version: '3.11'
 
       - name: Set up chart-testing
-        uses: helm/chart-testing-action@v2.4.0
+        uses: helm/chart-testing-action@v2.6.0
 
       - name: Run chart-testing (list-changed)
         id: list-changed

.github/workflows/global-run-codeql.yaml

@@ -27,6 +27,11 @@ jobs:
       - name: Checkout repository
         uses: actions/checkout@v3
 
+      - name: Install Go
+        uses: actions/setup-go@v4
+        with:
+          go-version-file: go.mod
+
       - name: Initialize CodeQL
         uses: github/codeql-action/init@v2
         with:

.github/workflows/go.yaml

@@ -51,7 +51,7 @@ jobs:
         uses: golangci/golangci-lint-action@v3
         with:
           only-new-issues: true
-          version: v1.51.1
+          version: v1.55.2
 
   generate:
     runs-on: ubuntu-latest

CODEOWNERS

@@ -1,2 +1,2 @@
 # actions-runner-controller maintainers
-* @mumoshu @toast-gear @actions/actions-runtime @nikola-jokic
+* @mumoshu @toast-gear @actions/actions-launch @nikola-jokic @rentziass

CONTRIBUTING.md

@@ -73,7 +73,7 @@ To make your development cycle faster, use the below command to update deploy an
 # Makefile
 VERSION=controller1 \
   RUNNER_TAG=runner1 \
-  make acceptance/pull acceptance/kind docker-build acceptance/load acceptance/deploy
+  make acceptance/pull acceptance/kind docker-buildx acceptance/load acceptance/deploy
 ```
 
 If you've already deployed actions-runner-controller and only want to recreate pods to use the newer image, you can run:

Dockerfile

@@ -1,5 +1,5 @@
 # Build the manager binary
-FROM --platform=$BUILDPLATFORM golang:1.20.7 as builder
+FROM --platform=$BUILDPLATFORM golang:1.22.1 as builder
 
 WORKDIR /workspace
 
@@ -38,6 +38,7 @@ RUN --mount=target=. \
   export GOOS=${TARGETOS} GOARCH=${TARGETARCH} GOARM=${TARGETVARIANT#v} && \
   go build -trimpath -ldflags="-s -w -X 'github.com/actions/actions-runner-controller/build.Version=${VERSION}' -X 'github.com/actions/actions-runner-controller/build.CommitSHA=${COMMIT_SHA}'" -o /out/manager main.go && \
   go build -trimpath -ldflags="-s -w -X 'github.com/actions/actions-runner-controller/build.Version=${VERSION}' -X 'github.com/actions/actions-runner-controller/build.CommitSHA=${COMMIT_SHA}'" -o /out/github-runnerscaleset-listener ./cmd/githubrunnerscalesetlistener && \
+  go build -trimpath -ldflags="-s -w -X 'github.com/actions/actions-runner-controller/build.Version=${VERSION}' -X 'github.com/actions/actions-runner-controller/build.CommitSHA=${COMMIT_SHA}'" -o /out/ghalistener ./cmd/ghalistener && \
   go build -trimpath -ldflags="-s -w" -o /out/github-webhook-server ./cmd/githubwebhookserver && \
   go build -trimpath -ldflags="-s -w" -o /out/actions-metrics-server ./cmd/actionsmetricsserver && \
   go build -trimpath -ldflags="-s -w" -o /out/sleep ./cmd/sleep
@@ -52,6 +53,7 @@ COPY --from=builder /out/manager .
 COPY --from=builder /out/github-webhook-server .
 COPY --from=builder /out/actions-metrics-server .
 COPY --from=builder /out/github-runnerscaleset-listener .
+COPY --from=builder /out/ghalistener .
 COPY --from=builder /out/sleep .
 
 USER 65532:65532

Makefile

@@ -6,7 +6,7 @@ endif
 DOCKER_USER ?= $(shell echo ${DOCKER_IMAGE_NAME} | cut -d / -f1)
 VERSION ?= dev
 COMMIT_SHA = $(shell git rev-parse HEAD)
-RUNNER_VERSION ?= 2.309.0
+RUNNER_VERSION ?= 2.315.0
 TARGETPLATFORM ?= $(shell arch)
 RUNNER_NAME ?= ${DOCKER_USER}/actions-runner
 RUNNER_TAG  ?= ${VERSION}
@@ -68,7 +68,7 @@ endif
 all: manager
 
 lint:
-	docker run --rm -v $(PWD):/app -w /app golangci/golangci-lint:v1.54.2 golangci-lint run
+	docker run --rm -v $(PWD):/app -w /app golangci/golangci-lint:v1.55.2 golangci-lint run
 
 GO_TEST_ARGS ?= -short
 
@@ -320,7 +320,7 @@ ifeq (, $(wildcard $(GOBIN)/controller-gen))
 	CONTROLLER_GEN_TMP_DIR=$$(mktemp -d) ;\
 	cd $$CONTROLLER_GEN_TMP_DIR ;\
 	go mod init tmp ;\
-	go install sigs.k8s.io/controller-tools/cmd/controller-gen@v0.11.3 ;\
+	go install sigs.k8s.io/controller-tools/cmd/controller-gen@v0.14.0 ;\
 	rm -rf $$CONTROLLER_GEN_TMP_DIR ;\
 	}
 endif

apis/actions.github.com/v1alpha1/ephemeralrunner_types.go

@@ -42,6 +42,10 @@ type EphemeralRunner struct {
 	Status EphemeralRunnerStatus `json:"status,omitempty"`
 }
 
+func (er *EphemeralRunner) IsDone() bool {
+	return er.Status.Phase == corev1.PodSucceeded || er.Status.Phase == corev1.PodFailed
+}
+
 // EphemeralRunnerSpec defines the desired state of EphemeralRunner
 type EphemeralRunnerSpec struct {
 	// INSERT ADDITIONAL SPEC FIELDS - desired state of cluster

apis/actions.github.com/v1alpha1/ephemeralrunnerset_types.go

@@ -24,6 +24,8 @@ import (
 type EphemeralRunnerSetSpec struct {
 	// Replicas is the number of desired EphemeralRunner resources in the k8s namespace.
 	Replicas int `json:"replicas,omitempty"`
+	// PatchID is the unique identifier for the patch issued by the listener app
+	PatchID int `json:"patchID"`
 
 	EphemeralRunnerSpec EphemeralRunnerSpec `json:"ephemeralRunnerSpec,omitempty"`
 }

apis/actions.github.com/v1alpha1/zz_generated.deepcopy.go

@@ -1,5 +1,4 @@
 //go:build !ignore_autogenerated
-// +build !ignore_autogenerated
 
 /*
 Copyright 2020 The actions-runner-controller authors.

apis/actions.summerwind.net/v1alpha1/runner_webhook.go

@@ -23,6 +23,7 @@ import (
 	ctrl "sigs.k8s.io/controller-runtime"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 	"sigs.k8s.io/controller-runtime/pkg/webhook"
+	"sigs.k8s.io/controller-runtime/pkg/webhook/admission"
 )
 
 // log is for logging in this package.
@@ -48,20 +49,20 @@ func (r *Runner) Default() {
 var _ webhook.Validator = &Runner{}
 
 // ValidateCreate implements webhook.Validator so a webhook will be registered for the type
-func (r *Runner) ValidateCreate() error {
+func (r *Runner) ValidateCreate() (admission.Warnings, error) {
 	runnerLog.Info("validate resource to be created", "name", r.Name)
-	return r.Validate()
+	return nil, r.Validate()
 }
 
 // ValidateUpdate implements webhook.Validator so a webhook will be registered for the type
-func (r *Runner) ValidateUpdate(old runtime.Object) error {
+func (r *Runner) ValidateUpdate(old runtime.Object) (admission.Warnings, error) {
 	runnerLog.Info("validate resource to be updated", "name", r.Name)
-	return r.Validate()
+	return nil, r.Validate()
 }
 
 // ValidateDelete implements webhook.Validator so a webhook will be registered for the type
-func (r *Runner) ValidateDelete() error {
-	return nil
+func (r *Runner) ValidateDelete() (admission.Warnings, error) {
+	return nil, nil
 }
 
 // Validate validates resource spec.

apis/actions.summerwind.net/v1alpha1/runnerdeployment_webhook.go

@@ -23,6 +23,7 @@ import (
 	ctrl "sigs.k8s.io/controller-runtime"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 	"sigs.k8s.io/controller-runtime/pkg/webhook"
+	"sigs.k8s.io/controller-runtime/pkg/webhook/admission"
 )
 
 // log is for logging in this package.
@@ -48,20 +49,20 @@ func (r *RunnerDeployment) Default() {
 var _ webhook.Validator = &RunnerDeployment{}
 
 // ValidateCreate implements webhook.Validator so a webhook will be registered for the type
-func (r *RunnerDeployment) ValidateCreate() error {
+func (r *RunnerDeployment) ValidateCreate() (admission.Warnings, error) {
 	runnerDeploymentLog.Info("validate resource to be created", "name", r.Name)
-	return r.Validate()
+	return nil, r.Validate()
 }
 
 // ValidateUpdate implements webhook.Validator so a webhook will be registered for the type
-func (r *RunnerDeployment) ValidateUpdate(old runtime.Object) error {
+func (r *RunnerDeployment) ValidateUpdate(old runtime.Object) (admission.Warnings, error) {
 	runnerDeploymentLog.Info("validate resource to be updated", "name", r.Name)
-	return r.Validate()
+	return nil, r.Validate()
 }
 
 // ValidateDelete implements webhook.Validator so a webhook will be registered for the type
-func (r *RunnerDeployment) ValidateDelete() error {
-	return nil
+func (r *RunnerDeployment) ValidateDelete() (admission.Warnings, error) {
+	return nil, nil
 }
 
 // Validate validates resource spec.

apis/actions.summerwind.net/v1alpha1/runnerreplicaset_webhook.go

@@ -23,6 +23,7 @@ import (
 	ctrl "sigs.k8s.io/controller-runtime"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 	"sigs.k8s.io/controller-runtime/pkg/webhook"
+	"sigs.k8s.io/controller-runtime/pkg/webhook/admission"
 )
 
 // log is for logging in this package.
@@ -48,20 +49,20 @@ func (r *RunnerReplicaSet) Default() {
 var _ webhook.Validator = &RunnerReplicaSet{}
 
 // ValidateCreate implements webhook.Validator so a webhook will be registered for the type
-func (r *RunnerReplicaSet) ValidateCreate() error {
+func (r *RunnerReplicaSet) ValidateCreate() (admission.Warnings, error) {
 	runnerReplicaSetLog.Info("validate resource to be created", "name", r.Name)
-	return r.Validate()
+	return nil, r.Validate()
 }
 
 // ValidateUpdate implements webhook.Validator so a webhook will be registered for the type
-func (r *RunnerReplicaSet) ValidateUpdate(old runtime.Object) error {
+func (r *RunnerReplicaSet) ValidateUpdate(old runtime.Object) (admission.Warnings, error) {
 	runnerReplicaSetLog.Info("validate resource to be updated", "name", r.Name)
-	return r.Validate()
+	return nil, r.Validate()
 }
 
 // ValidateDelete implements webhook.Validator so a webhook will be registered for the type
-func (r *RunnerReplicaSet) ValidateDelete() error {
-	return nil
+func (r *RunnerReplicaSet) ValidateDelete() (admission.Warnings, error) {
+	return nil, nil
 }
 
 // Validate validates resource spec.

apis/actions.summerwind.net/v1alpha1/zz_generated.deepcopy.go

@@ -1,5 +1,4 @@
 //go:build !ignore_autogenerated
-// +build !ignore_autogenerated
 
 /*
 Copyright 2020 The actions-runner-controller authors.

charts/actions-runner-controller/Chart.yaml

@@ -15,10 +15,10 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.23.5
+version: 0.23.7
 
 # Used as the default manager tag value when no tag property is provided in the values.yaml
-appVersion: 0.27.5
+appVersion: 0.27.6
 
 home: https://github.com/actions/actions-runner-controller
 

charts/actions-runner-controller/README.md

@@ -9,7 +9,7 @@ All additional docs are kept in the `docs/` folder, this README is solely for do
 > _Default values are the defaults set in the charts `values.yaml`, some properties have default configurations in the code for when the property is omitted or invalid_
 
 | Key                                                       | Description                                                                                                                               | Default                                                                                         |
-|----------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------|
+|-----------------------------------------------------------|-------------------------------------------------------------------------------------------------------------------------------------------|-------------------------------------------------------------------------------------------------|
 | `labels`                                                  | Set labels to apply to all resources in the chart                                                                                         |                                                                                                 |
 | `replicaCount`                                            | Set the number of controller pods                                                                                                         | 1                                                                                               |
 | `webhookPort`                                             | Set the containerPort for the webhook Pod                                                                                                 | 9443                                                                                            |
@@ -44,6 +44,7 @@ All additional docs are kept in the `docs/` folder, this README is solely for do
 | `image.pullPolicy`                                        | The pull policy of the controller image                                                                                                   | IfNotPresent                                                                                    |
 | `metrics.serviceMonitor.enable`                           | Deploy serviceMonitor kind for for use with prometheus-operator CRDs                                                                      | false                                                                                           |
 | `metrics.serviceMonitor.interval`                         | Configure the interval that Prometheus should scrap the controller's metrics                                                              | 1m                                                                                              |
+| `metrics.serviceMonitor.namespace                         | Namespace which Prometheus is running in                                                                                                  | `Release.Namespace` (the default namespace of the helm chart).                                  |
 | `metrics.serviceMonitor.timeout`                          | Configure the timeout the timeout of Prometheus scrapping.                                                                                | 30s                                                                                             |
 | `metrics.serviceAnnotations`                              | Set annotations for the provisioned metrics service resource                                                                              |                                                                                                 |
 | `metrics.port`                                            | Set port of metrics service                                                                                                               | 8443                                                                                            |
@@ -153,6 +154,7 @@ All additional docs are kept in the `docs/` folder, this README is solely for do
 | `actionsMetricsServer.ingress.ingressClassName`           | Set ingress class name                                                                                                                    |                                                                                                 |
 | `actionsMetrics.serviceMonitor.enable`                    | Deploy serviceMonitor kind for for use with prometheus-operator CRDs                                                                      | false                                                                                           |
 | `actionsMetrics.serviceMonitor.interval`                  | Configure the interval that Prometheus should scrap the controller's metrics                                                              | 1m                                                                                              |
+| `actionsMetrics.serviceMonitor.namespace`                 | Namespace which Prometheus is running in.                                                                                                 | `Release.Namespace` (the default namespace of the helm chart).                                  |
 | `actionsMetrics.serviceMonitor.timeout`                   | Configure the timeout the timeout of Prometheus scrapping.                                                                                | 30s                                                                                             |
 | `actionsMetrics.serviceAnnotations`                       | Set annotations for the provisioned actions metrics service resource                                                                      |                                                                                                 |
 | `actionsMetrics.port`                                     | Set port of actions metrics service                                                                                                       | 8443                                                                                            |

charts/actions-runner-controller/crds/actions.summerwind.dev_horizontalrunnerautoscalers.yaml

@@ -3,8 +3,7 @@ apiVersion: apiextensions.k8s.io/v1
 kind: CustomResourceDefinition
 metadata:
   annotations:
-    controller-gen.kubebuilder.io/version: v0.11.3
-  creationTimestamp: null
+    controller-gen.kubebuilder.io/version: v0.14.0
   name: horizontalrunnerautoscalers.actions.summerwind.dev
 spec:
   group: actions.summerwind.dev
@@ -36,10 +35,19 @@ spec:
           description: HorizontalRunnerAutoscaler is the Schema for the horizontalrunnerautoscaler API
           properties:
             apiVersion:
-              description: 'APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources'
+              description: |-
+                APIVersion defines the versioned schema of this representation of an object.
+                Servers should convert recognized schemas to the latest internal value, and
+                may reject unrecognized values.
+                More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
               type: string
             kind:
-              description: 'Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds'
+              description: |-
+                Kind is a string value representing the REST resource this object represents.
+                Servers may infer this from the endpoint the client submits requests to.
+                Cannot be updated.
+                In CamelCase.
+                More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
               type: string
             metadata:
               type: object
@@ -48,7 +56,9 @@ spec:
               properties:
                 capacityReservations:
                   items:
-                    description: CapacityReservation specifies the number of replicas temporarily added to the scale target until ExpirationTime.
+                    description: |-
+                      CapacityReservation specifies the number of replicas temporarily added
+                      to the scale target until ExpirationTime.
                     properties:
                       effectiveTime:
                         format: date-time
@@ -80,30 +90,46 @@ spec:
                   items:
                     properties:
                       repositoryNames:
-                        description: RepositoryNames is the list of repository names to be used for calculating the metric. For example, a repository name is the REPO part of `github.com/USER/REPO`.
+                        description: |-
+                          RepositoryNames is the list of repository names to be used for calculating the metric.
+                          For example, a repository name is the REPO part of `github.com/USER/REPO`.
                         items:
                           type: string
                         type: array
                       scaleDownAdjustment:
-                        description: ScaleDownAdjustment is the number of runners removed on scale-down. You can only specify either ScaleDownFactor or ScaleDownAdjustment.
+                        description: |-
+                          ScaleDownAdjustment is the number of runners removed on scale-down.
+                          You can only specify either ScaleDownFactor or ScaleDownAdjustment.
                         type: integer
                       scaleDownFactor:
-                        description: ScaleDownFactor is the multiplicative factor applied to the current number of runners used to determine how many pods should be removed.
+                        description: |-
+                          ScaleDownFactor is the multiplicative factor applied to the current number of runners used
+                          to determine how many pods should be removed.
                         type: string
                       scaleDownThreshold:
-                        description: ScaleDownThreshold is the percentage of busy runners less than which will trigger the hpa to scale the runners down.
+                        description: |-
+                          ScaleDownThreshold is the percentage of busy runners less than which will
+                          trigger the hpa to scale the runners down.
                         type: string
                       scaleUpAdjustment:
-                        description: ScaleUpAdjustment is the number of runners added on scale-up. You can only specify either ScaleUpFactor or ScaleUpAdjustment.
+                        description: |-
+                          ScaleUpAdjustment is the number of runners added on scale-up.
+                          You can only specify either ScaleUpFactor or ScaleUpAdjustment.
                         type: integer
                       scaleUpFactor:
-                        description: ScaleUpFactor is the multiplicative factor applied to the current number of runners used to determine how many pods should be added.
+                        description: |-
+                          ScaleUpFactor is the multiplicative factor applied to the current number of runners used
+                          to determine how many pods should be added.
                         type: string
                       scaleUpThreshold:
-                        description: ScaleUpThreshold is the percentage of busy runners greater than which will trigger the hpa to scale runners up.
+                        description: |-
+                          ScaleUpThreshold is the percentage of busy runners greater than which will
+                          trigger the hpa to scale runners up.
                         type: string
                       type:
-                        description: Type is the type of metric to be used for autoscaling. It can be TotalNumberOfQueuedAndInProgressWorkflowRuns or PercentageRunnersBusy.
+                        description: |-
+                          Type is the type of metric to be used for autoscaling.
+                          It can be TotalNumberOfQueuedAndInProgressWorkflowRuns or PercentageRunnersBusy.
                         type: string
                     type: object
                   type: array
@@ -111,7 +137,9 @@ spec:
                   description: MinReplicas is the minimum number of replicas the deployment is allowed to scale
                   type: integer
                 scaleDownDelaySecondsAfterScaleOut:
-                  description: ScaleDownDelaySecondsAfterScaleUp is the approximate delay for a scale down followed by a scale up Used to prevent flapping (down->up->down->... loop)
+                  description: |-
+                    ScaleDownDelaySecondsAfterScaleUp is the approximate delay for a scale down followed by a scale up
+                    Used to prevent flapping (down->up->down->... loop)
                   type: integer
                 scaleTargetRef:
                   description: ScaleTargetRef is the reference to scaled resource like RunnerDeployment
@@ -127,7 +155,18 @@ spec:
                       type: string
                   type: object
                 scaleUpTriggers:
-                  description: "ScaleUpTriggers is an experimental feature to increase the desired replicas by 1 on each webhook requested received by the webhookBasedAutoscaler. \n This feature requires you to also enable and deploy the webhookBasedAutoscaler onto your cluster. \n Note that the added runners remain until the next sync period at least, and they may or may not be used by GitHub Actions depending on the timing. They are intended to be used to gain \"resource slack\" immediately after you receive a webhook from GitHub, so that you can loosely expect MinReplicas runners to be always available."
+                  description: |-
+                    ScaleUpTriggers is an experimental feature to increase the desired replicas by 1
+                    on each webhook requested received by the webhookBasedAutoscaler.
+
+
+                    This feature requires you to also enable and deploy the webhookBasedAutoscaler onto your cluster.
+
+
+                    Note that the added runners remain until the next sync period at least,
+                    and they may or may not be used by GitHub Actions depending on the timing.
+                    They are intended to be used to gain "resource slack" immediately after you
+                    receive a webhook from GitHub, so that you can loosely expect MinReplicas runners to be always available.
                   items:
                     properties:
                       amount:
@@ -140,12 +179,18 @@ spec:
                             description: https://docs.github.com/en/actions/reference/events-that-trigger-workflows#check_run
                             properties:
                               names:
-                                description: Names is a list of GitHub Actions glob patterns. Any check_run event whose name matches one of patterns in the list can trigger autoscaling. Note that check_run name seem to equal to the job name you've defined in your actions workflow yaml file. So it is very likely that you can utilize this to trigger depending on the job.
+                                description: |-
+                                  Names is a list of GitHub Actions glob patterns.
+                                  Any check_run event whose name matches one of patterns in the list can trigger autoscaling.
+                                  Note that check_run name seem to equal to the job name you've defined in your actions workflow yaml file.
+                                  So it is very likely that you can utilize this to trigger depending on the job.
                                 items:
                                   type: string
                                 type: array
                               repositories:
-                                description: Repositories is a list of GitHub repositories. Any check_run event whose repository matches one of repositories in the list can trigger autoscaling.
+                                description: |-
+                                  Repositories is a list of GitHub repositories.
+                                  Any check_run event whose repository matches one of repositories in the list can trigger autoscaling.
                                 items:
                                   type: string
                                 type: array
@@ -170,7 +215,9 @@ spec:
                                 type: array
                             type: object
                           push:
-                            description: PushSpec is the condition for triggering scale-up on push event Also see https://docs.github.com/en/actions/reference/events-that-trigger-workflows#push
+                            description: |-
+                              PushSpec is the condition for triggering scale-up on push event
+                              Also see https://docs.github.com/en/actions/reference/events-that-trigger-workflows#push
                             type: object
                           workflowJob:
                             description: https://docs.github.com/en/developers/webhooks-and-events/webhooks/webhook-events-and-payloads#workflow_job
@@ -179,23 +226,33 @@ spec:
                     type: object
                   type: array
                 scheduledOverrides:
-                  description: ScheduledOverrides is the list of ScheduledOverride. It can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule. The earlier a scheduled override is, the higher it is prioritized.
+                  description: |-
+                    ScheduledOverrides is the list of ScheduledOverride.
+                    It can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule.
+                    The earlier a scheduled override is, the higher it is prioritized.
                   items:
-                    description: ScheduledOverride can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule. A schedule can optionally be recurring, so that the corresponding override happens every day, week, month, or year.
+                    description: |-
+                      ScheduledOverride can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule.
+                      A schedule can optionally be recurring, so that the corresponding override happens every day, week, month, or year.
                     properties:
                       endTime:
                         description: EndTime is the time at which the first override ends.
                         format: date-time
                         type: string
                       minReplicas:
-                        description: MinReplicas is the number of runners while overriding. If omitted, it doesn't override minReplicas.
+                        description: |-
+                          MinReplicas is the number of runners while overriding.
+                          If omitted, it doesn't override minReplicas.
                         minimum: 0
                         nullable: true
                         type: integer
                       recurrenceRule:
                         properties:
                           frequency:
-                            description: Frequency is the name of a predefined interval of each recurrence. The valid values are "Daily", "Weekly", "Monthly", and "Yearly". If empty, the corresponding override happens only once.
+                            description: |-
+                              Frequency is the name of a predefined interval of each recurrence.
+                              The valid values are "Daily", "Weekly", "Monthly", and "Yearly".
+                              If empty, the corresponding override happens only once.
                             enum:
                               - Daily
                               - Weekly
@@ -203,7 +260,9 @@ spec:
                               - Yearly
                             type: string
                           untilTime:
-                            description: UntilTime is the time of the final recurrence. If empty, the schedule recurs forever.
+                            description: |-
+                              UntilTime is the time of the final recurrence.
+                              If empty, the schedule recurs forever.
                             format: date-time
                             type: string
                         type: object
@@ -232,18 +291,24 @@ spec:
                     type: object
                   type: array
                 desiredReplicas:
-                  description: DesiredReplicas is the total number of desired, non-terminated and latest pods to be set for the primary RunnerSet This doesn't include outdated pods while upgrading the deployment and replacing the runnerset.
+                  description: |-
+                    DesiredReplicas is the total number of desired, non-terminated and latest pods to be set for the primary RunnerSet
+                    This doesn't include outdated pods while upgrading the deployment and replacing the runnerset.
                   type: integer
                 lastSuccessfulScaleOutTime:
                   format: date-time
                   nullable: true
                   type: string
                 observedGeneration:
-                  description: ObservedGeneration is the most recent generation observed for the target. It corresponds to e.g. RunnerDeployment's generation, which is updated on mutation by the API Server.
+                  description: |-
+                    ObservedGeneration is the most recent generation observed for the target. It corresponds to e.g.
+                    RunnerDeployment's generation, which is updated on mutation by the API Server.
                   format: int64
                   type: integer
                 scheduledOverridesSummary:
-                  description: ScheduledOverridesSummary is the summary of active and upcoming scheduled overrides to be shown in e.g. a column of a `kubectl get hra` output for observability.
+                  description: |-
+                    ScheduledOverridesSummary is the summary of active and upcoming scheduled overrides to be shown in e.g. a column of a `kubectl get hra` output
+                    for observability.
                   type: string
               type: object
           type: object

charts/actions-runner-controller/templates/actionsmetrics.servicemonitor.yaml.yml

@@ -1,4 +1,5 @@
 {{- if and .Values.actionsMetricsServer.enabled .Values.actionsMetrics.serviceMonitor.enable }}
+{{- $servicemonitornamespace := .Values.actionsMetrics.serviceMonitor.namespace | default .Release.Namespace }}
 apiVersion: monitoring.coreos.com/v1
 kind: ServiceMonitor
 metadata:
@@ -8,7 +9,7 @@ metadata:
     {{- toYaml . | nindent 4 }}
   {{- end }}
   name: {{ include "actions-runner-controller-actions-metrics-server.serviceMonitorName" . }}
-  namespace: {{ .Release.Namespace }}
+  namespace: {{ $servicemonitornamespace }}
 spec:
   endpoints:
     - path: /metrics

charts/actions-runner-controller/templates/githubwebhook.serviceMonitor.yaml

@@ -1,4 +1,5 @@
 {{- if and .Values.githubWebhookServer.enabled .Values.metrics.serviceMonitor.enable }}
+{{- $servicemonitornamespace := .Values.actionsMetrics.serviceMonitor.namespace | default .Release.Namespace }}
 apiVersion: monitoring.coreos.com/v1
 kind: ServiceMonitor
 metadata:
@@ -8,7 +9,7 @@ metadata:
     {{- toYaml . | nindent 4 }}
   {{- end }}
   name: {{ include "actions-runner-controller-github-webhook-server.serviceMonitorName" . }}
-  namespace: {{ .Release.Namespace }}
+  namespace: {{ $servicemonitornamespace }}
 spec:
   endpoints:
     - path: /metrics

charts/actions-runner-controller/values.yaml

@@ -111,6 +111,7 @@ metrics:
   serviceAnnotations: {}
   serviceMonitor:
     enable: false
+    namespace: ""
     timeout: 30s
     interval: 1m
   serviceMonitorLabels: {}
@@ -312,6 +313,7 @@ actionsMetrics:
   # to deploy the actions-metrics-server whose k8s service is referenced by the service monitor.
   serviceMonitor:
     enable: false
+    namespace: ""
     timeout: 30s
     interval: 1m
   serviceMonitorLabels: {}

charts/gha-runner-scale-set-controller/Chart.yaml

@@ -15,13 +15,13 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.6.1
+version: 0.9.1
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to
 # follow Semantic Versioning. They should reflect the version the application is using.
 # It is recommended to use it with quotes.
-appVersion: "0.6.1"
+appVersion: "0.9.1"
 
 home: https://github.com/actions/actions-runner-controller
 

charts/gha-runner-scale-set-controller/templates/NOTES.txt

@@ -2,4 +2,4 @@ Thank you for installing {{ .Chart.Name }}.
 
 Your release is named {{ .Release.Name }}.
 
-WARNING: value specified under image.pullPolicy will be ignored and no longer be applied to the listener pod spec as of gha-runner-scale-set-0.7.0. Please use the listenerTemplate in the gha-runner-scale-set chart to control the image pull policy of the listener.
+WARNING: Older version of the listener (githubrunnerscalesetlistener) is deprecated and will be removed in the future gha-runner-scale-set-0.10.0 release. If you are using environment variable override to force the old listener, please remove the environment variable and use the new listener (ghalistener) instead.

charts/gha-runner-scale-set-controller/templates/_helpers.tpl

@@ -48,7 +48,7 @@ app.kubernetes.io/version: {{ .Chart.AppVersion | quote }}
 app.kubernetes.io/part-of: gha-rs-controller
 app.kubernetes.io/managed-by: {{ .Release.Service }}
 {{- range $k, $v := .Values.labels }}
-{{ $k }}: {{ $v }}
+{{ $k }}: {{ $v | quote }}
 {{- end }}
 {{- end }}
 

charts/gha-runner-scale-set-controller/templates/deployment.yaml

@@ -27,6 +27,9 @@ spec:
         app.kubernetes.io/component: controller-manager
         app.kubernetes.io/version: {{ .Chart.Version }}
         {{- include "gha-runner-scale-set-controller.selectorLabels" . | nindent 8 }}
+        {{- with .Values.podLabels }}
+          {{- toYaml . | nindent 8 }}
+        {{- end }}
     spec:
       {{- with .Values.imagePullSecrets }}
       imagePullSecrets:
@@ -91,8 +94,6 @@ spec:
           valueFrom:
             fieldRef:
               fieldPath: metadata.namespace
-        - name: CONTROLLER_MANAGER_LISTENER_IMAGE_PULL_POLICY
-          value: "{{ .Values.image.pullPolicy | default "IfNotPresent" }}"
         {{- with .Values.env }}
           {{- if kindIs "slice" . }}
             {{- toYaml . | nindent 8 }}
@@ -109,10 +110,16 @@ spec:
         volumeMounts:
         - mountPath: /tmp
           name: tmp
+        {{- range .Values.volumeMounts }}
+        - {{ toYaml . | nindent 10 }}
+        {{- end }}
       terminationGracePeriodSeconds: 10
       volumes:
       - name: tmp
         emptyDir: {}
+      {{- range .Values.volumes }}
+      - {{ toYaml . | nindent 8 }}
+      {{- end }}
       {{- with .Values.nodeSelector }}
       nodeSelector:
         {{- toYaml . | nindent 8 }}
@@ -121,6 +128,10 @@ spec:
       affinity:
         {{- toYaml . | nindent 8 }}
       {{- end }}
+      {{- with .Values.topologySpreadConstraints }}
+      topologySpreadConstraints:
+        {{- toYaml . | nindent 8 }}
+      {{- end }}
       {{- with .Values.tolerations }}
       tolerations:
         {{- toYaml . | nindent 8 }}

charts/gha-runner-scale-set-controller/tests/template_test.go

@@ -345,6 +345,7 @@ func TestTemplate_ControllerDeployment_Defaults(t *testing.T) {
 
 	assert.Len(t, deployment.Spec.Template.Spec.NodeSelector, 0)
 	assert.Nil(t, deployment.Spec.Template.Spec.Affinity)
+	assert.Len(t, deployment.Spec.Template.Spec.TopologySpreadConstraints, 0)
 	assert.Len(t, deployment.Spec.Template.Spec.Tolerations, 0)
 
 	managerImage := "ghcr.io/actions/gha-runner-scale-set-controller:dev"
@@ -368,14 +369,12 @@ func TestTemplate_ControllerDeployment_Defaults(t *testing.T) {
 	}
 	assert.ElementsMatch(t, expectedArgs, deployment.Spec.Template.Spec.Containers[0].Args)
 
-	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 3)
+	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 2)
 	assert.Equal(t, "CONTROLLER_MANAGER_CONTAINER_IMAGE", deployment.Spec.Template.Spec.Containers[0].Env[0].Name)
 	assert.Equal(t, managerImage, deployment.Spec.Template.Spec.Containers[0].Env[0].Value)
 
 	assert.Equal(t, "CONTROLLER_MANAGER_POD_NAMESPACE", deployment.Spec.Template.Spec.Containers[0].Env[1].Name)
 	assert.Equal(t, "metadata.namespace", deployment.Spec.Template.Spec.Containers[0].Env[1].ValueFrom.FieldRef.FieldPath)
-	assert.Equal(t, "CONTROLLER_MANAGER_LISTENER_IMAGE_PULL_POLICY", deployment.Spec.Template.Spec.Containers[0].Env[2].Name)
-	assert.Equal(t, "IfNotPresent", deployment.Spec.Template.Spec.Containers[0].Env[2].Value) // default value. Needs to align with controllers/actions.github.com/resourcebuilder.go
 
 	assert.Empty(t, deployment.Spec.Template.Spec.Containers[0].Resources)
 	assert.Nil(t, deployment.Spec.Template.Spec.Containers[0].SecurityContext)
@@ -406,6 +405,8 @@ func TestTemplate_ControllerDeployment_Customize(t *testing.T) {
 		SetValues: map[string]string{
 			"labels.foo":                   "bar",
 			"labels.github":                "actions",
+			"labels.team":                  "GitHub Team",
+			"labels.teamMail":              "team@github.com",
 			"replicaCount":                 "1",
 			"image.pullPolicy":             "Always",
 			"image.tag":                    "dev",
@@ -424,10 +425,17 @@ func TestTemplate_ControllerDeployment_Customize(t *testing.T) {
 			"tolerations[0].key":           "foo",
 			"affinity.nodeAffinity.requiredDuringSchedulingIgnoredDuringExecution.nodeSelectorTerms[0].matchExpressions[0].key":      "foo",
 			"affinity.nodeAffinity.requiredDuringSchedulingIgnoredDuringExecution.nodeSelectorTerms[0].matchExpressions[0].operator": "bar",
+			"topologySpreadConstraints[0].labelSelector.matchLabels.foo":                                                             "bar",
+			"topologySpreadConstraints[0].maxSkew":                                                                                   "1",
+			"topologySpreadConstraints[0].topologyKey":                                                                               "foo",
 			"priorityClassName":         "test-priority-class",
 			"flags.updateStrategy":      "eventual",
 			"flags.logLevel":            "info",
 			"flags.logFormat":           "json",
+			"volumes[0].name":           "customMount",
+			"volumes[0].configMap.name": "my-configmap",
+			"volumeMounts[0].name":      "customMount",
+			"volumeMounts[0].mountPath": "/my/mount/path",
 		},
 		KubectlOptions: k8s.NewKubectlOptions("", "", namespaceName),
 	}
@@ -447,6 +455,8 @@ func TestTemplate_ControllerDeployment_Customize(t *testing.T) {
 	assert.Equal(t, "gha-rs-controller", deployment.Labels["app.kubernetes.io/part-of"])
 	assert.Equal(t, "bar", deployment.Labels["foo"])
 	assert.Equal(t, "actions", deployment.Labels["github"])
+	assert.Equal(t, "GitHub Team", deployment.Labels["team"])
+	assert.Equal(t, "team@github.com", deployment.Labels["teamMail"])
 
 	assert.Equal(t, int32(1), *deployment.Spec.Replicas)
 
@@ -459,8 +469,8 @@ func TestTemplate_ControllerDeployment_Customize(t *testing.T) {
 	assert.Equal(t, "bar", deployment.Spec.Template.Annotations["foo"])
 	assert.Equal(t, "manager", deployment.Spec.Template.Annotations["kubectl.kubernetes.io/default-container"])
 
-	assert.Equal(t, "ENV_VAR_NAME_1", deployment.Spec.Template.Spec.Containers[0].Env[3].Name)
-	assert.Equal(t, "ENV_VAR_VALUE_1", deployment.Spec.Template.Spec.Containers[0].Env[3].Value)
+	assert.Equal(t, "ENV_VAR_NAME_1", deployment.Spec.Template.Spec.Containers[0].Env[2].Name)
+	assert.Equal(t, "ENV_VAR_VALUE_1", deployment.Spec.Template.Spec.Containers[0].Env[2].Value)
 
 	assert.Len(t, deployment.Spec.Template.Spec.ImagePullSecrets, 1)
 	assert.Equal(t, "dockerhub", deployment.Spec.Template.Spec.ImagePullSecrets[0].Name)
@@ -468,9 +478,11 @@ func TestTemplate_ControllerDeployment_Customize(t *testing.T) {
 	assert.Equal(t, int64(1000), *deployment.Spec.Template.Spec.SecurityContext.FSGroup)
 	assert.Equal(t, "test-priority-class", deployment.Spec.Template.Spec.PriorityClassName)
 	assert.Equal(t, int64(10), *deployment.Spec.Template.Spec.TerminationGracePeriodSeconds)
-	assert.Len(t, deployment.Spec.Template.Spec.Volumes, 1)
+	assert.Len(t, deployment.Spec.Template.Spec.Volumes, 2)
 	assert.Equal(t, "tmp", deployment.Spec.Template.Spec.Volumes[0].Name)
-	assert.NotNil(t, 10, deployment.Spec.Template.Spec.Volumes[0].EmptyDir)
+	assert.NotNil(t, deployment.Spec.Template.Spec.Volumes[0].EmptyDir)
+	assert.Equal(t, "customMount", deployment.Spec.Template.Spec.Volumes[1].Name)
+	assert.Equal(t, "my-configmap", deployment.Spec.Template.Spec.Volumes[1].ConfigMap.Name)
 
 	assert.Len(t, deployment.Spec.Template.Spec.NodeSelector, 1)
 	assert.Equal(t, "bar", deployment.Spec.Template.Spec.NodeSelector["foo"])
@@ -479,6 +491,11 @@ func TestTemplate_ControllerDeployment_Customize(t *testing.T) {
 	assert.Equal(t, "foo", deployment.Spec.Template.Spec.Affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution.NodeSelectorTerms[0].MatchExpressions[0].Key)
 	assert.Equal(t, "bar", string(deployment.Spec.Template.Spec.Affinity.NodeAffinity.RequiredDuringSchedulingIgnoredDuringExecution.NodeSelectorTerms[0].MatchExpressions[0].Operator))
 
+	assert.Len(t, deployment.Spec.Template.Spec.TopologySpreadConstraints, 1)
+	assert.Equal(t, "bar", deployment.Spec.Template.Spec.TopologySpreadConstraints[0].LabelSelector.MatchLabels["foo"])
+	assert.Equal(t, int32(1), deployment.Spec.Template.Spec.TopologySpreadConstraints[0].MaxSkew)
+	assert.Equal(t, "foo", deployment.Spec.Template.Spec.TopologySpreadConstraints[0].TopologyKey)
+
 	assert.Len(t, deployment.Spec.Template.Spec.Tolerations, 1)
 	assert.Equal(t, "foo", deployment.Spec.Template.Spec.Tolerations[0].Key)
 
@@ -505,25 +522,25 @@ func TestTemplate_ControllerDeployment_Customize(t *testing.T) {
 
 	assert.ElementsMatch(t, expectArgs, deployment.Spec.Template.Spec.Containers[0].Args)
 
-	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 4)
+	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 3)
 	assert.Equal(t, "CONTROLLER_MANAGER_CONTAINER_IMAGE", deployment.Spec.Template.Spec.Containers[0].Env[0].Name)
 	assert.Equal(t, managerImage, deployment.Spec.Template.Spec.Containers[0].Env[0].Value)
 
 	assert.Equal(t, "CONTROLLER_MANAGER_POD_NAMESPACE", deployment.Spec.Template.Spec.Containers[0].Env[1].Name)
 	assert.Equal(t, "metadata.namespace", deployment.Spec.Template.Spec.Containers[0].Env[1].ValueFrom.FieldRef.FieldPath)
-	assert.Equal(t, "CONTROLLER_MANAGER_LISTENER_IMAGE_PULL_POLICY", deployment.Spec.Template.Spec.Containers[0].Env[2].Name)
-	assert.Equal(t, "Always", deployment.Spec.Template.Spec.Containers[0].Env[2].Value) // default value. Needs to align with controllers/actions.github.com/resourcebuilder.go
 
-	assert.Equal(t, "ENV_VAR_NAME_1", deployment.Spec.Template.Spec.Containers[0].Env[3].Name)
-	assert.Equal(t, "ENV_VAR_VALUE_1", deployment.Spec.Template.Spec.Containers[0].Env[3].Value)
+	assert.Equal(t, "ENV_VAR_NAME_1", deployment.Spec.Template.Spec.Containers[0].Env[2].Name)
+	assert.Equal(t, "ENV_VAR_VALUE_1", deployment.Spec.Template.Spec.Containers[0].Env[2].Value)
 
 	assert.Equal(t, "500m", deployment.Spec.Template.Spec.Containers[0].Resources.Limits.Cpu().String())
 	assert.True(t, *deployment.Spec.Template.Spec.Containers[0].SecurityContext.RunAsNonRoot)
 	assert.Equal(t, int64(1000), *deployment.Spec.Template.Spec.Containers[0].SecurityContext.RunAsUser)
 
-	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].VolumeMounts, 1)
+	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].VolumeMounts, 2)
 	assert.Equal(t, "tmp", deployment.Spec.Template.Spec.Containers[0].VolumeMounts[0].Name)
 	assert.Equal(t, "/tmp", deployment.Spec.Template.Spec.Containers[0].VolumeMounts[0].MountPath)
+	assert.Equal(t, "customMount", deployment.Spec.Template.Spec.Containers[0].VolumeMounts[1].Name)
+	assert.Equal(t, "/my/mount/path", deployment.Spec.Template.Spec.Containers[0].VolumeMounts[1].MountPath)
 }
 
 func TestTemplate_EnableLeaderElectionRole(t *testing.T) {
@@ -737,6 +754,7 @@ func TestTemplate_ControllerDeployment_WatchSingleNamespace(t *testing.T) {
 
 	assert.Len(t, deployment.Spec.Template.Spec.NodeSelector, 0)
 	assert.Nil(t, deployment.Spec.Template.Spec.Affinity)
+	assert.Len(t, deployment.Spec.Template.Spec.TopologySpreadConstraints, 0)
 	assert.Len(t, deployment.Spec.Template.Spec.Tolerations, 0)
 
 	managerImage := "ghcr.io/actions/gha-runner-scale-set-controller:dev"
@@ -762,14 +780,12 @@ func TestTemplate_ControllerDeployment_WatchSingleNamespace(t *testing.T) {
 
 	assert.ElementsMatch(t, expectedArgs, deployment.Spec.Template.Spec.Containers[0].Args)
 
-	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 3)
+	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 2)
 	assert.Equal(t, "CONTROLLER_MANAGER_CONTAINER_IMAGE", deployment.Spec.Template.Spec.Containers[0].Env[0].Name)
 	assert.Equal(t, managerImage, deployment.Spec.Template.Spec.Containers[0].Env[0].Value)
 
 	assert.Equal(t, "CONTROLLER_MANAGER_POD_NAMESPACE", deployment.Spec.Template.Spec.Containers[0].Env[1].Name)
 	assert.Equal(t, "metadata.namespace", deployment.Spec.Template.Spec.Containers[0].Env[1].ValueFrom.FieldRef.FieldPath)
-	assert.Equal(t, "CONTROLLER_MANAGER_LISTENER_IMAGE_PULL_POLICY", deployment.Spec.Template.Spec.Containers[0].Env[2].Name)
-	assert.Equal(t, "IfNotPresent", deployment.Spec.Template.Spec.Containers[0].Env[2].Value) // default value. Needs to align with controllers/actions.github.com/resourcebuilder.go
 
 	assert.Empty(t, deployment.Spec.Template.Spec.Containers[0].Resources)
 	assert.Nil(t, deployment.Spec.Template.Spec.Containers[0].SecurityContext)
@@ -812,17 +828,17 @@ func TestTemplate_ControllerContainerEnvironmentVariables(t *testing.T) {
 	assert.Equal(t, namespaceName, deployment.Namespace)
 	assert.Equal(t, "test-arc-gha-rs-controller", deployment.Name)
 
-	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 7)
-	assert.Equal(t, "ENV_VAR_NAME_1", deployment.Spec.Template.Spec.Containers[0].Env[3].Name)
-	assert.Equal(t, "ENV_VAR_VALUE_1", deployment.Spec.Template.Spec.Containers[0].Env[3].Value)
-	assert.Equal(t, "ENV_VAR_NAME_2", deployment.Spec.Template.Spec.Containers[0].Env[4].Name)
-	assert.Equal(t, "secret-name", deployment.Spec.Template.Spec.Containers[0].Env[4].ValueFrom.SecretKeyRef.Name)
-	assert.Equal(t, "ENV_VAR_NAME_2", deployment.Spec.Template.Spec.Containers[0].Env[4].ValueFrom.SecretKeyRef.Key)
-	assert.True(t, *deployment.Spec.Template.Spec.Containers[0].Env[4].ValueFrom.SecretKeyRef.Optional)
-	assert.Equal(t, "ENV_VAR_NAME_3", deployment.Spec.Template.Spec.Containers[0].Env[5].Name)
-	assert.Empty(t, deployment.Spec.Template.Spec.Containers[0].Env[5].Value)
-	assert.Equal(t, "ENV_VAR_NAME_4", deployment.Spec.Template.Spec.Containers[0].Env[6].Name)
-	assert.Empty(t, deployment.Spec.Template.Spec.Containers[0].Env[6].ValueFrom)
+	assert.Len(t, deployment.Spec.Template.Spec.Containers[0].Env, 6)
+	assert.Equal(t, "ENV_VAR_NAME_1", deployment.Spec.Template.Spec.Containers[0].Env[2].Name)
+	assert.Equal(t, "ENV_VAR_VALUE_1", deployment.Spec.Template.Spec.Containers[0].Env[2].Value)
+	assert.Equal(t, "ENV_VAR_NAME_2", deployment.Spec.Template.Spec.Containers[0].Env[3].Name)
+	assert.Equal(t, "secret-name", deployment.Spec.Template.Spec.Containers[0].Env[3].ValueFrom.SecretKeyRef.Name)
+	assert.Equal(t, "ENV_VAR_NAME_2", deployment.Spec.Template.Spec.Containers[0].Env[3].ValueFrom.SecretKeyRef.Key)
+	assert.True(t, *deployment.Spec.Template.Spec.Containers[0].Env[3].ValueFrom.SecretKeyRef.Optional)
+	assert.Equal(t, "ENV_VAR_NAME_3", deployment.Spec.Template.Spec.Containers[0].Env[4].Name)
+	assert.Empty(t, deployment.Spec.Template.Spec.Containers[0].Env[4].Value)
+	assert.Equal(t, "ENV_VAR_NAME_4", deployment.Spec.Template.Spec.Containers[0].Env[5].Name)
+	assert.Empty(t, deployment.Spec.Template.Spec.Containers[0].Env[5].ValueFrom)
 }
 
 func TestTemplate_WatchSingleNamespace_NotCreateManagerClusterRole(t *testing.T) {

charts/gha-runner-scale-set-controller/values.yaml

@@ -41,6 +41,8 @@ serviceAccount:
 
 podAnnotations: {}
 
+podLabels: {}
+
 podSecurityContext: {}
 # fsGroup: 2000
 
@@ -70,6 +72,12 @@ tolerations: []
 
 affinity: {}
 
+topologySpreadConstraints: []
+
+# Mount volumes in the container.
+volumes: []
+volumeMounts: []
+
 # Leverage a PriorityClass to ensure your pods survive resource shortages
 # ref: https://kubernetes.io/docs/concepts/configuration/pod-priority-preemption/
 # PriorityClass: system-cluster-critical
@@ -100,7 +108,7 @@ flags:
 
   ## Defines how the controller should handle upgrades while having running jobs.
   ##
-  ## The srategies available are:
+  ## The strategies available are:
   ## - "immediate": (default) The controller will immediately apply the change causing the
   ##   recreation of the listener and ephemeral runner set. This can lead to an
   ##   overprovisioning of runners, if there are pending / running jobs. This should not

charts/gha-runner-scale-set/Chart.yaml

@@ -15,18 +15,18 @@ type: application
 # This is the chart version. This version number should be incremented each time you make changes
 # to the chart and its templates, including the app version.
 # Versions are expected to follow Semantic Versioning (https://semver.org/)
-version: 0.6.1
+version: 0.9.1
 
 # This is the version number of the application being deployed. This version number should be
 # incremented each time you make changes to the application. Versions are not expected to
 # follow Semantic Versioning. They should reflect the version the application is using.
 # It is recommended to use it with quotes.
-appVersion: "0.6.1"
+appVersion: "0.9.1"
 
-home: https://github.com/actions/dev-arc
+home: https://github.com/actions/actions-runner-controller
 
 sources:
-  - "https://github.com/actions/dev-arc"
+  - "https://github.com/actions/actions-runner-controller"
 
 maintainers:
   - name: actions

charts/gha-runner-scale-set/templates/_helpers.tpl

@@ -99,7 +99,7 @@ volumeMounts:
 image: docker:dind
 args:
   - dockerd
-  - --host=unix:///run/docker/docker.sock
+  - --host=unix:///var/run/docker.sock
   - --group=$(DOCKER_GROUP_GID)
 env:
   - name: DOCKER_GROUP_GID
@@ -110,7 +110,7 @@ volumeMounts:
   - name: work
     mountPath: /home/runner/_work
   - name: dind-sock
-    mountPath: /run/docker
+    mountPath: /var/run
   - name: dind-externals
     mountPath: /home/runner/externals
 {{- end }}
@@ -223,7 +223,7 @@ env:
     {{- end }}
     {{- if $setDockerHost }}
   - name: DOCKER_HOST
-    value: unix:///run/docker/docker.sock
+    value: unix:///var/run/docker.sock
     {{- end }}
     {{- if $setRunnerWaitDocker }}
   - name: RUNNER_WAIT_FOR_DOCKER_IN_SECONDS
@@ -264,8 +264,7 @@ volumeMounts:
     {{- end }}
     {{- if $mountDindCert }}
   - name: dind-sock
-    mountPath: /run/docker
-    readOnly: true
+    mountPath: /var/run
     {{- end }}
     {{- if $mountGitHubServerTLS }}
   - name: github-server-tls-cert
@@ -385,6 +384,9 @@ volumeMounts:
     {{- $setNodeExtraCaCerts = 1 }}
     {{- $setRunnerUpdateCaCerts = 1 }}
   {{- end }}
+
+  {{- $mountGitHubServerTLS := 0 }}
+  {{- if or $container.env $setNodeExtraCaCerts $setRunnerUpdateCaCerts }}
   env:
     {{- with $container.env }}
       {{- range $i, $env := . }}
@@ -405,10 +407,12 @@ volumeMounts:
     - name: RUNNER_UPDATE_CA_CERTS
       value: "1"
     {{- end }}
-    {{- $mountGitHubServerTLS := 0 }}
     {{- if $tlsConfig.runnerMountPath }}
       {{- $mountGitHubServerTLS = 1 }}
     {{- end }}
+  {{- end }}
+
+  {{- if or $container.volumeMounts $mountGitHubServerTLS }}
   volumeMounts:
     {{- with $container.volumeMounts }}
       {{- range $i, $volMount := . }}
@@ -423,6 +427,7 @@ volumeMounts:
       mountPath: {{ clean (print $tlsConfig.runnerMountPath "/" $tlsConfig.certificateFrom.configMapKeyRef.key) }}
       subPath: {{ $tlsConfig.certificateFrom.configMapKeyRef.key }}
     {{- end }}
+  {{- end}}
 {{- end }}
 {{- end }}
 {{- end }}
@@ -520,13 +525,13 @@ volumeMounts:
     {{- end }}
   {{- end }}
   {{- if and (eq $multiNamespacesCounter 0) (eq $singleNamespaceCounter 0) }}
-    {{- fail "No gha-rs-controller deployment found using label (app.kubernetes.io/part-of=gha-rs-controller). Consider setting controllerServiceAccount.name in values.yaml to be explicit if you think the discovery is wrong." }}
+    {{- fail "No gha-rs-controller deployment found using label (app.kubernetes.io/part-of=gha-rs-controller). Consider setting controllerServiceAccount.namespace in values.yaml to be explicit if you think the discovery is wrong." }}
   {{- end }}
   {{- if and (gt $multiNamespacesCounter 0) (gt $singleNamespaceCounter 0) }}
-    {{- fail "Found both gha-rs-controller installed with flags.watchSingleNamespace set and unset in cluster, this is not supported. Consider setting controllerServiceAccount.name in values.yaml to be explicit if you think the discovery is wrong." }}
+    {{- fail "Found both gha-rs-controller installed with flags.watchSingleNamespace set and unset in cluster, this is not supported. Consider setting controllerServiceAccount.namespace in values.yaml to be explicit if you think the discovery is wrong." }}
   {{- end }}
   {{- if gt $multiNamespacesCounter 1 }}
-    {{- fail "More than one gha-rs-controller deployment found using label (app.kubernetes.io/part-of=gha-rs-controller). Consider setting controllerServiceAccount.name in values.yaml to be explicit if you think the discovery is wrong." }}
+    {{- fail "More than one gha-rs-controller deployment found using label (app.kubernetes.io/part-of=gha-rs-controller). Consider setting controllerServiceAccount.namespace in values.yaml to be explicit if you think the discovery is wrong." }}
   {{- end }}
   {{- if eq $multiNamespacesCounter 1 }}
     {{- with $controllerDeployment.metadata }}
@@ -539,11 +544,11 @@ volumeMounts:
         {{- $managerServiceAccountNamespace = (get $controllerDeployment.metadata.labels "actions.github.com/controller-service-account-namespace") }}
       {{- end }}
     {{- else }}
-      {{- fail "No gha-rs-controller deployment that watch this namespace found using label (actions.github.com/controller-watch-single-namespace). Consider setting controllerServiceAccount.name in values.yaml to be explicit if you think the discovery is wrong." }}
+      {{- fail "No gha-rs-controller deployment that watch this namespace found using label (actions.github.com/controller-watch-single-namespace). Consider setting controllerServiceAccount.namespace in values.yaml to be explicit if you think the discovery is wrong." }}
     {{- end }}
   {{- end }}
   {{- if eq $managerServiceAccountNamespace "" }}
-    {{- fail "No service account namespace found for gha-rs-controller deployment using label (actions.github.com/controller-service-account-namespace), consider setting controllerServiceAccount.name in values.yaml to be explicit if you think the discovery is wrong." }}
+    {{- fail "No service account namespace found for gha-rs-controller deployment using label (actions.github.com/controller-service-account-namespace), consider setting controllerServiceAccount.namespace in values.yaml to be explicit if you think the discovery is wrong." }}
   {{- end }}
 {{- $managerServiceAccountNamespace }}
 {{- end }}

charts/gha-runner-scale-set/templates/autoscalingrunnerset.yaml

@@ -13,6 +13,7 @@ metadata:
     app.kubernetes.io/component: "autoscaling-runner-set"
     {{- include "gha-runner-scale-set.labels" . | nindent 4 }}
   annotations:
+    actions.github.com/values-hash: {{ toJson .Values | sha256sum | trunc 63 }}
     {{- $containerMode := .Values.containerMode }}
     {{- if not (kindIs "string" .Values.githubConfigSecret) }}
     actions.github.com/cleanup-github-secret-name: {{ include "gha-runner-scale-set.githubsecret" . }}

charts/gha-runner-scale-set/templates/kube_mode_serviceaccount.yaml

@@ -5,6 +5,12 @@ kind: ServiceAccount
 metadata:
   name: {{ include "gha-runner-scale-set.kubeModeServiceAccountName" . }}
   namespace: {{ .Release.Namespace }}
+  {{- if .Values.containerMode.kubernetesModeServiceAccount }}
+  {{- with .Values.containerMode.kubernetesModeServiceAccount.annotations }}
+  annotations:
+  {{- toYaml . | nindent 4 }}
+  {{- end }}
+  {{- end }}
   finalizers:
     - actions.github.com/cleanup-protection
   labels:

charts/gha-runner-scale-set/tests/template_test.go

@@ -742,6 +742,37 @@ func TestTemplateRenderedAutoScalingRunnerSet_DinD_ExtraInitContainers(t *testin
 	assert.Equal(t, "ls", ars.Spec.Template.Spec.InitContainers[2].Command[0], "InitContainers[2] Command[0] should be ls")
 }
 
+func TestTemplateRenderedKubernetesModeServiceAccountAnnotations(t *testing.T) {
+	t.Parallel()
+
+	// Path to the helm chart we will test
+	helmChartPath, err := filepath.Abs("../../gha-runner-scale-set")
+	require.NoError(t, err)
+
+	testValuesPath, err := filepath.Abs("../tests/values_kubernetes_mode_service_account_annotations.yaml")
+	require.NoError(t, err)
+
+	releaseName := "test-runners"
+	namespaceName := "test-" + strings.ToLower(random.UniqueId())
+
+	options := &helm.Options{
+		Logger: logger.Discard,
+		SetValues: map[string]string{
+			"controllerServiceAccount.name":      "arc",
+			"controllerServiceAccount.namespace": "arc-system",
+		},
+		ValuesFiles:    []string{testValuesPath},
+		KubectlOptions: k8s.NewKubectlOptions("", "", namespaceName),
+	}
+
+	output := helm.RenderTemplate(t, options, helmChartPath, releaseName, []string{"templates/kube_mode_serviceaccount.yaml"})
+
+	var sa corev1.ServiceAccount
+	helm.UnmarshalK8SYaml(t, output, &sa)
+
+	assert.Equal(t, "arn:aws:iam::123456789012:role/sample-role", sa.Annotations["eks.amazonaws.com/role-arn"], "Annotations should be arn:aws:iam::123456789012:role/sample-role")
+}
+
 func TestTemplateRenderedAutoScalingRunnerSet_DinD_ExtraVolumes(t *testing.T) {
 	t.Parallel()
 
@@ -869,7 +900,7 @@ func TestTemplateRenderedAutoScalingRunnerSet_EnableDinD(t *testing.T) {
 	assert.Equal(t, "ghcr.io/actions/actions-runner:latest", ars.Spec.Template.Spec.Containers[0].Image)
 	assert.Len(t, ars.Spec.Template.Spec.Containers[0].Env, 2, "The runner container should have 2 env vars, DOCKER_HOST and RUNNER_WAIT_FOR_DOCKER_IN_SECONDS")
 	assert.Equal(t, "DOCKER_HOST", ars.Spec.Template.Spec.Containers[0].Env[0].Name)
-	assert.Equal(t, "unix:///run/docker/docker.sock", ars.Spec.Template.Spec.Containers[0].Env[0].Value)
+	assert.Equal(t, "unix:///var/run/docker.sock", ars.Spec.Template.Spec.Containers[0].Env[0].Value)
 	assert.Equal(t, "RUNNER_WAIT_FOR_DOCKER_IN_SECONDS", ars.Spec.Template.Spec.Containers[0].Env[1].Name)
 	assert.Equal(t, "120", ars.Spec.Template.Spec.Containers[0].Env[1].Value)
 
@@ -879,8 +910,7 @@ func TestTemplateRenderedAutoScalingRunnerSet_EnableDinD(t *testing.T) {
 	assert.False(t, ars.Spec.Template.Spec.Containers[0].VolumeMounts[0].ReadOnly)
 
 	assert.Equal(t, "dind-sock", ars.Spec.Template.Spec.Containers[0].VolumeMounts[1].Name)
-	assert.Equal(t, "/run/docker", ars.Spec.Template.Spec.Containers[0].VolumeMounts[1].MountPath)
-	assert.True(t, ars.Spec.Template.Spec.Containers[0].VolumeMounts[1].ReadOnly)
+	assert.Equal(t, "/var/run", ars.Spec.Template.Spec.Containers[0].VolumeMounts[1].MountPath)
 
 	assert.Equal(t, "dind", ars.Spec.Template.Spec.Containers[1].Name)
 	assert.Equal(t, "docker:dind", ars.Spec.Template.Spec.Containers[1].Image)
@@ -890,7 +920,7 @@ func TestTemplateRenderedAutoScalingRunnerSet_EnableDinD(t *testing.T) {
 	assert.Equal(t, "/home/runner/_work", ars.Spec.Template.Spec.Containers[1].VolumeMounts[0].MountPath)
 
 	assert.Equal(t, "dind-sock", ars.Spec.Template.Spec.Containers[1].VolumeMounts[1].Name)
-	assert.Equal(t, "/run/docker", ars.Spec.Template.Spec.Containers[1].VolumeMounts[1].MountPath)
+	assert.Equal(t, "/var/run", ars.Spec.Template.Spec.Containers[1].VolumeMounts[1].MountPath)
 
 	assert.Equal(t, "dind-externals", ars.Spec.Template.Spec.Containers[1].VolumeMounts[2].Name)
 	assert.Equal(t, "/home/runner/externals", ars.Spec.Template.Spec.Containers[1].VolumeMounts[2].MountPath)
@@ -1986,3 +2016,130 @@ func TestTemplateRenderedAutoscalingRunnerSetAnnotation_KubernetesModeCleanup(t
 		assert.Equal(t, value, autoscalingRunnerSet.Annotations[annotation], fmt.Sprintf("Annotation %q does not match the expected value", annotation))
 	}
 }
+
+func TestRunnerContainerEnvNotEmptyMap(t *testing.T) {
+	t.Parallel()
+
+	// Path to the helm chart we will test
+	helmChartPath, err := filepath.Abs("../../gha-runner-scale-set")
+	require.NoError(t, err)
+
+	testValuesPath, err := filepath.Abs("../tests/values.yaml")
+	require.NoError(t, err)
+
+	releaseName := "test-runners"
+	namespaceName := "test-" + strings.ToLower(random.UniqueId())
+
+	options := &helm.Options{
+		Logger:         logger.Discard,
+		ValuesFiles:    []string{testValuesPath},
+		KubectlOptions: k8s.NewKubectlOptions("", "", namespaceName),
+	}
+
+	output := helm.RenderTemplate(t, options, helmChartPath, releaseName, []string{"templates/autoscalingrunnerset.yaml"})
+	type testModel struct {
+		Spec struct {
+			Template struct {
+				Spec struct {
+					Containers []map[string]any `yaml:"containers"`
+				} `yaml:"spec"`
+			} `yaml:"template"`
+		} `yaml:"spec"`
+	}
+
+	var m testModel
+	helm.UnmarshalK8SYaml(t, output, &m)
+	_, ok := m.Spec.Template.Spec.Containers[0]["env"]
+	assert.False(t, ok, "env should not be set")
+}
+
+func TestRunnerContainerVolumeNotEmptyMap(t *testing.T) {
+	t.Parallel()
+
+	// Path to the helm chart we will test
+	helmChartPath, err := filepath.Abs("../../gha-runner-scale-set")
+	require.NoError(t, err)
+
+	testValuesPath, err := filepath.Abs("../tests/values.yaml")
+	require.NoError(t, err)
+
+	releaseName := "test-runners"
+	namespaceName := "test-" + strings.ToLower(random.UniqueId())
+
+	options := &helm.Options{
+		Logger:         logger.Discard,
+		ValuesFiles:    []string{testValuesPath},
+		KubectlOptions: k8s.NewKubectlOptions("", "", namespaceName),
+	}
+
+	output := helm.RenderTemplate(t, options, helmChartPath, releaseName, []string{"templates/autoscalingrunnerset.yaml"})
+	type testModel struct {
+		Spec struct {
+			Template struct {
+				Spec struct {
+					Containers []map[string]any `yaml:"containers"`
+				} `yaml:"spec"`
+			} `yaml:"template"`
+		} `yaml:"spec"`
+	}
+
+	var m testModel
+	helm.UnmarshalK8SYaml(t, output, &m)
+	_, ok := m.Spec.Template.Spec.Containers[0]["volumeMounts"]
+	assert.False(t, ok, "volumeMounts should not be set")
+}
+
+func TestAutoscalingRunnerSetAnnotationValuesHash(t *testing.T) {
+	t.Parallel()
+
+	const valuesHash = "actions.github.com/values-hash"
+
+	// Path to the helm chart we will test
+	helmChartPath, err := filepath.Abs("../../gha-runner-scale-set")
+	require.NoError(t, err)
+
+	releaseName := "test-runners"
+	namespaceName := "test-" + strings.ToLower(random.UniqueId())
+
+	options := &helm.Options{
+		Logger: logger.Discard,
+		SetValues: map[string]string{
+			"githubConfigUrl":                    "https://github.com/actions",
+			"githubConfigSecret.github_token":    "gh_token12345",
+			"controllerServiceAccount.name":      "arc",
+			"controllerServiceAccount.namespace": "arc-system",
+		},
+		KubectlOptions: k8s.NewKubectlOptions("", "", namespaceName),
+	}
+
+	output := helm.RenderTemplate(t, options, helmChartPath, releaseName, []string{"templates/autoscalingrunnerset.yaml"})
+
+	var autoscalingRunnerSet v1alpha1.AutoscalingRunnerSet
+	helm.UnmarshalK8SYaml(t, output, &autoscalingRunnerSet)
+
+	firstHash := autoscalingRunnerSet.Annotations["actions.github.com/values-hash"]
+	assert.NotEmpty(t, firstHash)
+	assert.LessOrEqual(t, len(firstHash), 63)
+
+	helmChartPath, err = filepath.Abs("../../gha-runner-scale-set")
+	require.NoError(t, err)
+
+	options = &helm.Options{
+		Logger: logger.Discard,
+		SetValues: map[string]string{
+			"githubConfigUrl":                    "https://github.com/actions",
+			"githubConfigSecret.github_token":    "gh_token1234567890",
+			"controllerServiceAccount.name":      "arc",
+			"controllerServiceAccount.namespace": "arc-system",
+		},
+		KubectlOptions: k8s.NewKubectlOptions("", "", namespaceName),
+	}
+
+	output = helm.RenderTemplate(t, options, helmChartPath, releaseName, []string{"templates/autoscalingrunnerset.yaml"})
+
+	helm.UnmarshalK8SYaml(t, output, &autoscalingRunnerSet)
+	secondHash := autoscalingRunnerSet.Annotations[valuesHash]
+	assert.NotEmpty(t, secondHash)
+	assert.NotEqual(t, firstHash, secondHash)
+	assert.LessOrEqual(t, len(secondHash), 63)
+}

charts/gha-runner-scale-set/tests/values_kubernetes_mode_service_account_annotations.yaml

@@ -0,0 +1,8 @@
+githubConfigUrl: https://github.com/actions/actions-runner-controller
+githubConfigSecret:
+  github_token: test
+containerMode:
+  type: kubernetes
+  kubernetesModeServiceAccount:
+    annotations:
+      eks.amazonaws.com/role-arn: arn:aws:iam::123456789012:role/sample-role

charts/gha-runner-scale-set/values.yaml

@@ -39,7 +39,8 @@ githubConfigSecret:
 ## maxRunners is the max number of runners the autoscaling runner set will scale up to.
 # maxRunners: 5
 
-## minRunners is the min number of runners the autoscaling runner set will scale down to.
+## minRunners is the min number of idle runners. The target number of runners created will be
+## calculated as a sum of minRunners and the number of jobs assigned to the scale set.
 # minRunners: 0
 
 # runnerGroup: "default"
@@ -84,8 +85,10 @@ githubConfigSecret:
 #     resources:
 #       requests:
 #         storage: 1Gi
+#   kubernetesModeServiceAccount:
+#     annotations:
 
-## template is the PodSpec for each listener Pod
+## listenerTemplate is the PodSpec for each listener Pod
 ## For reference: https://kubernetes.io/docs/reference/kubernetes-api/workload-resources/pod-v1/#PodSpec
 # listenerTemplate:
 #   spec:
@@ -122,18 +125,17 @@ template:
   ##       command: ["/home/runner/run.sh"]
   ##       env:
   ##         - name: DOCKER_HOST
-  ##           value: unix:///run/docker/docker.sock
+  ##           value: unix:///var/run/docker.sock
   ##       volumeMounts:
   ##         - name: work
   ##           mountPath: /home/runner/_work
   ##         - name: dind-sock
-  ##           mountPath: /run/docker
-  ##           readOnly: true
+  ##           mountPath: /var/run
   ##     - name: dind
   ##       image: docker:dind
   ##       args:
   ##         - dockerd
-  ##         - --host=unix:///run/docker/docker.sock
+  ##         - --host=unix:///var/run/docker.sock
   ##         - --group=$(DOCKER_GROUP_GID)
   ##       env:
   ##         - name: DOCKER_GROUP_GID
@@ -144,7 +146,7 @@ template:
   ##         - name: work
   ##           mountPath: /home/runner/_work
   ##         - name: dind-sock
-  ##           mountPath: /run/docker
+  ##           mountPath: /var/run
   ##         - name: dind-externals
   ##           mountPath: /home/runner/externals
   ##     volumes:

cmd/ghalistener/app/app.go

@@ -0,0 +1,137 @@
+package app
+
+import (
+	"context"
+	"errors"
+	"fmt"
+
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/config"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/listener"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/metrics"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/worker"
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/go-logr/logr"
+	"golang.org/x/sync/errgroup"
+)
+
+// App is responsible for initializing required components and running the app.
+type App struct {
+	// configured fields
+	config config.Config
+	logger logr.Logger
+
+	// initialized fields
+	listener Listener
+	worker   Worker
+	metrics  metrics.ServerPublisher
+}
+
+//go:generate mockery --name Listener --output ./mocks --outpkg mocks --case underscore
+type Listener interface {
+	Listen(ctx context.Context, handler listener.Handler) error
+}
+
+//go:generate mockery --name Worker --output ./mocks --outpkg mocks --case underscore
+type Worker interface {
+	HandleJobStarted(ctx context.Context, jobInfo *actions.JobStarted) error
+	HandleDesiredRunnerCount(ctx context.Context, count int, jobsCompleted int) (int, error)
+}
+
+func New(config config.Config) (*App, error) {
+	app := &App{
+		config: config,
+	}
+
+	ghConfig, err := actions.ParseGitHubConfigFromURL(config.ConfigureUrl)
+	if err != nil {
+		return nil, fmt.Errorf("failed to parse GitHub config from URL: %w", err)
+	}
+
+	{
+		logger, err := config.Logger()
+		if err != nil {
+			return nil, fmt.Errorf("failed to create logger: %w", err)
+		}
+		app.logger = logger.WithName("listener-app")
+	}
+
+	actionsClient, err := config.ActionsClient(app.logger)
+	if err != nil {
+		return nil, fmt.Errorf("failed to create actions client: %w", err)
+	}
+
+	if config.MetricsAddr != "" {
+		app.metrics = metrics.NewExporter(metrics.ExporterConfig{
+			ScaleSetName:      config.EphemeralRunnerSetName,
+			ScaleSetNamespace: config.EphemeralRunnerSetNamespace,
+			Enterprise:        ghConfig.Enterprise,
+			Organization:      ghConfig.Organization,
+			Repository:        ghConfig.Repository,
+			ServerAddr:        config.MetricsAddr,
+			ServerEndpoint:    config.MetricsEndpoint,
+		})
+	}
+
+	worker, err := worker.New(
+		worker.Config{
+			EphemeralRunnerSetNamespace: config.EphemeralRunnerSetNamespace,
+			EphemeralRunnerSetName:      config.EphemeralRunnerSetName,
+			MaxRunners:                  config.MaxRunners,
+			MinRunners:                  config.MinRunners,
+		},
+		worker.WithLogger(app.logger.WithName("worker")),
+	)
+	if err != nil {
+		return nil, fmt.Errorf("failed to create new kubernetes worker: %w", err)
+	}
+	app.worker = worker
+
+	listener, err := listener.New(listener.Config{
+		Client:     actionsClient,
+		ScaleSetID: app.config.RunnerScaleSetId,
+		MinRunners: app.config.MinRunners,
+		MaxRunners: app.config.MaxRunners,
+		Logger:     app.logger.WithName("listener"),
+		Metrics:    app.metrics,
+	})
+	if err != nil {
+		return nil, fmt.Errorf("failed to create new listener: %w", err)
+	}
+	app.listener = listener
+
+	app.logger.Info("app initialized")
+
+	return app, nil
+}
+
+func (app *App) Run(ctx context.Context) error {
+	var errs []error
+	if app.worker == nil {
+		errs = append(errs, fmt.Errorf("worker not initialized"))
+	}
+	if app.listener == nil {
+		errs = append(errs, fmt.Errorf("listener not initialized"))
+	}
+	if err := errors.Join(errs...); err != nil {
+		return fmt.Errorf("app not initialized: %w", err)
+	}
+
+	g, ctx := errgroup.WithContext(ctx)
+	metricsCtx, cancelMetrics := context.WithCancelCause(ctx)
+
+	g.Go(func() error {
+		app.logger.Info("Starting listener")
+		listnerErr := app.listener.Listen(ctx, app.worker)
+		cancelMetrics(fmt.Errorf("Listener exited: %w", listnerErr))
+		return listnerErr
+	})
+
+	if app.metrics != nil {
+		g.Go(func() error {
+			app.logger.Info("Starting metrics server")
+			return app.metrics.ListenAndServe(metricsCtx)
+		})
+	}
+
+	return g.Wait()
+}

cmd/ghalistener/app/app_test.go

@@ -0,0 +1,85 @@
+package app
+
+import (
+	"context"
+	"errors"
+	"testing"
+
+	appmocks "github.com/actions/actions-runner-controller/cmd/ghalistener/app/mocks"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/listener"
+	metricsMocks "github.com/actions/actions-runner-controller/cmd/ghalistener/metrics/mocks"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/worker"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/mock"
+)
+
+func TestApp_Run(t *testing.T) {
+	t.Parallel()
+
+	t.Run("ListenerWorkerGuard", func(t *testing.T) {
+		invalidApps := []*App{
+			{},
+			{worker: &worker.Worker{}},
+			{listener: &listener.Listener{}},
+		}
+
+		for _, app := range invalidApps {
+			assert.Error(t, app.Run(context.Background()))
+		}
+	})
+
+	t.Run("ExitsOnListenerError", func(t *testing.T) {
+		listener := appmocks.NewListener(t)
+		worker := appmocks.NewWorker(t)
+
+		listener.On("Listen", mock.Anything, mock.Anything).Return(errors.New("listener error")).Once()
+
+		app := &App{
+			listener: listener,
+			worker:   worker,
+		}
+
+		err := app.Run(context.Background())
+		assert.Error(t, err)
+	})
+
+	t.Run("ExitsOnListenerNil", func(t *testing.T) {
+		listener := appmocks.NewListener(t)
+		worker := appmocks.NewWorker(t)
+
+		listener.On("Listen", mock.Anything, mock.Anything).Return(nil).Once()
+
+		app := &App{
+			listener: listener,
+			worker:   worker,
+		}
+
+		err := app.Run(context.Background())
+		assert.NoError(t, err)
+	})
+
+	t.Run("CancelListenerOnMetricsServerError", func(t *testing.T) {
+		listener := appmocks.NewListener(t)
+		worker := appmocks.NewWorker(t)
+		metrics := metricsMocks.NewServerPublisher(t)
+		ctx := context.Background()
+
+		listener.On("Listen", mock.Anything, mock.Anything).Run(func(args mock.Arguments) {
+			ctx := args.Get(0).(context.Context)
+			go func() {
+				<-ctx.Done()
+			}()
+		}).Return(nil).Once()
+
+		metrics.On("ListenAndServe", mock.Anything).Return(errors.New("metrics server error")).Once()
+
+		app := &App{
+			listener: listener,
+			worker:   worker,
+			metrics:  metrics,
+		}
+
+		err := app.Run(ctx)
+		assert.Error(t, err)
+	})
+}

cmd/ghalistener/app/mocks/listener.go

@@ -0,0 +1,43 @@
+// Code generated by mockery v2.36.1. DO NOT EDIT.
+
+package mocks
+
+import (
+	context "context"
+
+	listener "github.com/actions/actions-runner-controller/cmd/ghalistener/listener"
+	mock "github.com/stretchr/testify/mock"
+)
+
+// Listener is an autogenerated mock type for the Listener type
+type Listener struct {
+	mock.Mock
+}
+
+// Listen provides a mock function with given fields: ctx, handler
+func (_m *Listener) Listen(ctx context.Context, handler listener.Handler) error {
+	ret := _m.Called(ctx, handler)
+
+	var r0 error
+	if rf, ok := ret.Get(0).(func(context.Context, listener.Handler) error); ok {
+		r0 = rf(ctx, handler)
+	} else {
+		r0 = ret.Error(0)
+	}
+
+	return r0
+}
+
+// NewListener creates a new instance of Listener. It also registers a testing interface on the mock and a cleanup function to assert the mocks expectations.
+// The first argument is typically a *testing.T value.
+func NewListener(t interface {
+	mock.TestingT
+	Cleanup(func())
+}) *Listener {
+	mock := &Listener{}
+	mock.Mock.Test(t)
+
+	t.Cleanup(func() { mock.AssertExpectations(t) })
+
+	return mock
+}

cmd/ghalistener/app/mocks/worker.go

@@ -0,0 +1,68 @@
+// Code generated by mockery v2.36.1. DO NOT EDIT.
+
+package mocks
+
+import (
+	actions "github.com/actions/actions-runner-controller/github/actions"
+
+	context "context"
+
+	mock "github.com/stretchr/testify/mock"
+)
+
+// Worker is an autogenerated mock type for the Worker type
+type Worker struct {
+	mock.Mock
+}
+
+// HandleDesiredRunnerCount provides a mock function with given fields: ctx, count, acquireCount
+func (_m *Worker) HandleDesiredRunnerCount(ctx context.Context, count int, acquireCount int) (int, error) {
+	ret := _m.Called(ctx, count, acquireCount)
+
+	var r0 int
+	var r1 error
+	if rf, ok := ret.Get(0).(func(context.Context, int, int) (int, error)); ok {
+		return rf(ctx, count, acquireCount)
+	}
+	if rf, ok := ret.Get(0).(func(context.Context, int, int) int); ok {
+		r0 = rf(ctx, count, acquireCount)
+	} else {
+		r0 = ret.Get(0).(int)
+	}
+
+	if rf, ok := ret.Get(1).(func(context.Context, int, int) error); ok {
+		r1 = rf(ctx, count, acquireCount)
+	} else {
+		r1 = ret.Error(1)
+	}
+
+	return r0, r1
+}
+
+// HandleJobStarted provides a mock function with given fields: ctx, jobInfo
+func (_m *Worker) HandleJobStarted(ctx context.Context, jobInfo *actions.JobStarted) error {
+	ret := _m.Called(ctx, jobInfo)
+
+	var r0 error
+	if rf, ok := ret.Get(0).(func(context.Context, *actions.JobStarted) error); ok {
+		r0 = rf(ctx, jobInfo)
+	} else {
+		r0 = ret.Error(0)
+	}
+
+	return r0
+}
+
+// NewWorker creates a new instance of Worker. It also registers a testing interface on the mock and a cleanup function to assert the mocks expectations.
+// The first argument is typically a *testing.T value.
+func NewWorker(t interface {
+	mock.TestingT
+	Cleanup(func())
+}) *Worker {
+	mock := &Worker{}
+	mock.Mock.Test(t)
+
+	t.Cleanup(func() { mock.AssertExpectations(t) })
+
+	return mock
+}

cmd/ghalistener/config/config.go

@@ -0,0 +1,161 @@
+package config
+
+import (
+	"crypto/x509"
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/url"
+	"os"
+
+	"github.com/actions/actions-runner-controller/build"
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/actions/actions-runner-controller/logging"
+	"github.com/go-logr/logr"
+	"golang.org/x/net/http/httpproxy"
+)
+
+type Config struct {
+	ConfigureUrl                string `json:"configureUrl"`
+	AppID                       int64  `json:"appID"`
+	AppInstallationID           int64  `json:"appInstallationID"`
+	AppPrivateKey               string `json:"appPrivateKey"`
+	Token                       string `json:"token"`
+	EphemeralRunnerSetNamespace string `json:"ephemeralRunnerSetNamespace"`
+	EphemeralRunnerSetName      string `json:"ephemeralRunnerSetName"`
+	MaxRunners                  int    `json:"maxRunners"`
+	MinRunners                  int    `json:"minRunners"`
+	RunnerScaleSetId            int    `json:"runnerScaleSetId"`
+	RunnerScaleSetName          string `json:"runnerScaleSetName"`
+	ServerRootCA                string `json:"serverRootCA"`
+	LogLevel                    string `json:"logLevel"`
+	LogFormat                   string `json:"logFormat"`
+	MetricsAddr                 string `json:"metricsAddr"`
+	MetricsEndpoint             string `json:"metricsEndpoint"`
+}
+
+func Read(path string) (Config, error) {
+	f, err := os.Open(path)
+	if err != nil {
+		return Config{}, err
+	}
+	defer f.Close()
+
+	var config Config
+	if err := json.NewDecoder(f).Decode(&config); err != nil {
+		return Config{}, fmt.Errorf("failed to decode config: %w", err)
+	}
+
+	if err := config.validate(); err != nil {
+		return Config{}, fmt.Errorf("failed to validate config: %w", err)
+	}
+
+	return config, nil
+}
+
+func (c *Config) validate() error {
+	if len(c.ConfigureUrl) == 0 {
+		return fmt.Errorf("GitHubConfigUrl is not provided")
+	}
+
+	if len(c.EphemeralRunnerSetNamespace) == 0 || len(c.EphemeralRunnerSetName) == 0 {
+		return fmt.Errorf("EphemeralRunnerSetNamespace '%s' or EphemeralRunnerSetName '%s' is missing", c.EphemeralRunnerSetNamespace, c.EphemeralRunnerSetName)
+	}
+
+	if c.RunnerScaleSetId == 0 {
+		return fmt.Errorf("RunnerScaleSetId '%d' is missing", c.RunnerScaleSetId)
+	}
+
+	if c.MaxRunners < c.MinRunners {
+		return fmt.Errorf("MinRunners '%d' cannot be greater than MaxRunners '%d'", c.MinRunners, c.MaxRunners)
+	}
+
+	hasToken := len(c.Token) > 0
+	hasPrivateKeyConfig := c.AppID > 0 && c.AppPrivateKey != ""
+
+	if !hasToken && !hasPrivateKeyConfig {
+		return fmt.Errorf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(c.Token), c.AppID, c.AppInstallationID, len(c.AppPrivateKey))
+	}
+
+	if hasToken && hasPrivateKeyConfig {
+		return fmt.Errorf("only one GitHub auth method supported at a time. Have both PAT and App auth: token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(c.Token), c.AppID, c.AppInstallationID, len(c.AppPrivateKey))
+	}
+
+	return nil
+}
+
+func (c *Config) Logger() (logr.Logger, error) {
+	logLevel := string(logging.LogLevelDebug)
+	if c.LogLevel != "" {
+		logLevel = c.LogLevel
+	}
+
+	logFormat := string(logging.LogFormatText)
+	if c.LogFormat != "" {
+		logFormat = c.LogFormat
+	}
+
+	logger, err := logging.NewLogger(logLevel, logFormat)
+	if err != nil {
+		return logr.Logger{}, fmt.Errorf("NewLogger failed: %w", err)
+	}
+
+	return logger, nil
+}
+
+func (c *Config) ActionsClient(logger logr.Logger, clientOptions ...actions.ClientOption) (*actions.Client, error) {
+	var creds actions.ActionsAuth
+	switch c.Token {
+	case "":
+		creds.AppCreds = &actions.GitHubAppAuth{
+			AppID:             c.AppID,
+			AppInstallationID: c.AppInstallationID,
+			AppPrivateKey:     c.AppPrivateKey,
+		}
+	default:
+		creds.Token = c.Token
+	}
+
+	options := append([]actions.ClientOption{
+		actions.WithLogger(logger),
+	}, clientOptions...)
+
+	if c.ServerRootCA != "" {
+		systemPool, err := x509.SystemCertPool()
+		if err != nil {
+			return nil, fmt.Errorf("failed to load system cert pool: %w", err)
+		}
+		pool := systemPool.Clone()
+		ok := pool.AppendCertsFromPEM([]byte(c.ServerRootCA))
+		if !ok {
+			return nil, fmt.Errorf("failed to parse root certificate")
+		}
+
+		options = append(options, actions.WithRootCAs(pool))
+	}
+
+	proxyFunc := httpproxy.FromEnvironment().ProxyFunc()
+	options = append(options, actions.WithProxy(func(req *http.Request) (*url.URL, error) {
+		return proxyFunc(req.URL)
+	}))
+
+	client, err := actions.NewClient(c.ConfigureUrl, &creds, options...)
+	if err != nil {
+		return nil, fmt.Errorf("failed to create actions client: %w", err)
+	}
+
+	client.SetUserAgent(actions.UserAgentInfo{
+		Version:    build.Version,
+		CommitSHA:  build.CommitSHA,
+		ScaleSetID: c.RunnerScaleSetId,
+		HasProxy:   hasProxy(),
+		Subsystem:  "ghalistener",
+	})
+
+	return client, nil
+}
+
+func hasProxy() bool {
+	proxyFunc := httpproxy.FromEnvironment().ProxyFunc()
+	return proxyFunc != nil
+}

cmd/ghalistener/config/config_client_test.go

@@ -0,0 +1,161 @@
+package config_test
+
+import (
+	"context"
+	"crypto/tls"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"path/filepath"
+	"testing"
+
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/config"
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/actions/actions-runner-controller/github/actions/testserver"
+	"github.com/go-logr/logr"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+func TestCustomerServerRootCA(t *testing.T) {
+	ctx := context.Background()
+	certsFolder := filepath.Join(
+		"../../../",
+		"github",
+		"actions",
+		"testdata",
+	)
+	certPath := filepath.Join(certsFolder, "server.crt")
+	keyPath := filepath.Join(certsFolder, "server.key")
+
+	serverCalledSuccessfully := false
+
+	server := testserver.NewUnstarted(t, http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		serverCalledSuccessfully = true
+		w.WriteHeader(http.StatusOK)
+		w.Write([]byte(`{"count": 0}`))
+	}))
+	cert, err := tls.LoadX509KeyPair(certPath, keyPath)
+	require.NoError(t, err)
+
+	server.TLS = &tls.Config{Certificates: []tls.Certificate{cert}}
+	server.StartTLS()
+
+	var certsString string
+	rootCA, err := os.ReadFile(filepath.Join(certsFolder, "rootCA.crt"))
+	require.NoError(t, err)
+	certsString = string(rootCA)
+
+	intermediate, err := os.ReadFile(filepath.Join(certsFolder, "intermediate.pem"))
+	require.NoError(t, err)
+	certsString = certsString + string(intermediate)
+
+	config := config.Config{
+		ConfigureUrl: server.ConfigURLForOrg("myorg"),
+		ServerRootCA: certsString,
+		Token:        "token",
+	}
+
+	client, err := config.ActionsClient(logr.Discard())
+	require.NoError(t, err)
+	_, err = client.GetRunnerScaleSet(ctx, 1, "test")
+	require.NoError(t, err)
+	assert.True(t, serverCalledSuccessfully)
+}
+
+func TestProxySettings(t *testing.T) {
+	t.Run("http", func(t *testing.T) {
+		wentThroughProxy := false
+
+		proxy := httptest.NewServer(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+			wentThroughProxy = true
+		}))
+		t.Cleanup(func() {
+			proxy.Close()
+		})
+
+		prevProxy := os.Getenv("http_proxy")
+		os.Setenv("http_proxy", proxy.URL)
+		defer os.Setenv("http_proxy", prevProxy)
+
+		config := config.Config{
+			ConfigureUrl: "https://github.com/org/repo",
+			Token:        "token",
+		}
+
+		client, err := config.ActionsClient(logr.Discard())
+		require.NoError(t, err)
+
+		req, err := http.NewRequest(http.MethodGet, "http://example.com", nil)
+		require.NoError(t, err)
+		_, err = client.Do(req)
+		require.NoError(t, err)
+
+		assert.True(t, wentThroughProxy)
+	})
+
+	t.Run("https", func(t *testing.T) {
+		wentThroughProxy := false
+
+		proxy := httptest.NewServer(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+			wentThroughProxy = true
+		}))
+		t.Cleanup(func() {
+			proxy.Close()
+		})
+
+		prevProxy := os.Getenv("https_proxy")
+		os.Setenv("https_proxy", proxy.URL)
+		defer os.Setenv("https_proxy", prevProxy)
+
+		config := config.Config{
+			ConfigureUrl: "https://github.com/org/repo",
+			Token:        "token",
+		}
+
+		client, err := config.ActionsClient(logr.Discard(), actions.WithRetryMax(0))
+		require.NoError(t, err)
+
+		req, err := http.NewRequest(http.MethodGet, "https://example.com", nil)
+		require.NoError(t, err)
+
+		_, err = client.Do(req)
+		// proxy doesn't support https
+		assert.Error(t, err)
+		assert.True(t, wentThroughProxy)
+	})
+
+	t.Run("no_proxy", func(t *testing.T) {
+		wentThroughProxy := false
+
+		proxy := httptest.NewServer(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+			wentThroughProxy = true
+		}))
+		t.Cleanup(func() {
+			proxy.Close()
+		})
+
+		prevProxy := os.Getenv("http_proxy")
+		os.Setenv("http_proxy", proxy.URL)
+		defer os.Setenv("http_proxy", prevProxy)
+
+		prevNoProxy := os.Getenv("no_proxy")
+		os.Setenv("no_proxy", "example.com")
+		defer os.Setenv("no_proxy", prevNoProxy)
+
+		config := config.Config{
+			ConfigureUrl: "https://github.com/org/repo",
+			Token:        "token",
+		}
+
+		client, err := config.ActionsClient(logr.Discard())
+		require.NoError(t, err)
+
+		req, err := http.NewRequest(http.MethodGet, "http://example.com", nil)
+		require.NoError(t, err)
+
+		_, err = client.Do(req)
+		require.NoError(t, err)
+		assert.False(t, wentThroughProxy)
+	})
+}

cmd/ghalistener/config/config_test.go

@@ -0,0 +1,92 @@
+package config
+
+import (
+	"fmt"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestConfigValidationMinMax(t *testing.T) {
+	config := &Config{
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+		MinRunners:                  5,
+		MaxRunners:                  2,
+		Token:                       "token",
+	}
+	err := config.validate()
+	assert.ErrorContains(t, err, "MinRunners '5' cannot be greater than MaxRunners '2", "Expected error about MinRunners > MaxRunners")
+}
+
+func TestConfigValidationMissingToken(t *testing.T) {
+	config := &Config{
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+	err := config.validate()
+	expectedError := fmt.Sprintf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
+	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
+}
+
+func TestConfigValidationAppKey(t *testing.T) {
+	config := &Config{
+		AppID:                       1,
+		AppInstallationID:           10,
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+	err := config.validate()
+	expectedError := fmt.Sprintf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
+	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
+}
+
+func TestConfigValidationOnlyOneTypeOfCredentials(t *testing.T) {
+	config := &Config{
+		AppID:                       1,
+		AppInstallationID:           10,
+		AppPrivateKey:               "asdf",
+		Token:                       "asdf",
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+	err := config.validate()
+	expectedError := fmt.Sprintf("only one GitHub auth method supported at a time. Have both PAT and App auth: token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
+	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
+}
+
+func TestConfigValidation(t *testing.T) {
+	config := &Config{
+		ConfigureUrl:                "https://github.com/actions",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+		MinRunners:                  1,
+		MaxRunners:                  5,
+		Token:                       "asdf",
+	}
+
+	err := config.validate()
+
+	assert.NoError(t, err, "Expected no error")
+}
+
+func TestConfigValidationConfigUrl(t *testing.T) {
+	config := &Config{
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+
+	err := config.validate()
+
+	assert.ErrorContains(t, err, "GitHubConfigUrl is not provided", "Expected error about missing ConfigureUrl")
+}

cmd/ghalistener/listener/listener.go

@@ -0,0 +1,437 @@
+package listener
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"net/http"
+	"os"
+	"time"
+
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/metrics"
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/go-logr/logr"
+	"github.com/google/uuid"
+)
+
+const (
+	sessionCreationMaxRetries = 10
+)
+
+// message types
+const (
+	messageTypeJobAvailable = "JobAvailable"
+	messageTypeJobAssigned  = "JobAssigned"
+	messageTypeJobStarted   = "JobStarted"
+	messageTypeJobCompleted = "JobCompleted"
+)
+
+//go:generate mockery --name Client --output ./mocks --outpkg mocks --case underscore
+type Client interface {
+	GetAcquirableJobs(ctx context.Context, runnerScaleSetId int) (*actions.AcquirableJobList, error)
+	CreateMessageSession(ctx context.Context, runnerScaleSetId int, owner string) (*actions.RunnerScaleSetSession, error)
+	GetMessage(ctx context.Context, messageQueueUrl, messageQueueAccessToken string, lastMessageId int64, maxCapacity int) (*actions.RunnerScaleSetMessage, error)
+	DeleteMessage(ctx context.Context, messageQueueUrl, messageQueueAccessToken string, messageId int64) error
+	AcquireJobs(ctx context.Context, runnerScaleSetId int, messageQueueAccessToken string, requestIds []int64) ([]int64, error)
+	RefreshMessageSession(ctx context.Context, runnerScaleSetId int, sessionId *uuid.UUID) (*actions.RunnerScaleSetSession, error)
+	DeleteMessageSession(ctx context.Context, runnerScaleSetId int, sessionId *uuid.UUID) error
+}
+
+type Config struct {
+	Client     Client
+	ScaleSetID int
+	MinRunners int
+	MaxRunners int
+	Logger     logr.Logger
+	Metrics    metrics.Publisher
+}
+
+func (c *Config) Validate() error {
+	if c.Client == nil {
+		return errors.New("client is required")
+	}
+	if c.ScaleSetID == 0 {
+		return errors.New("scaleSetID is required")
+	}
+	if c.MinRunners < 0 {
+		return errors.New("minRunners must be greater than or equal to 0")
+	}
+	if c.MaxRunners < 0 {
+		return errors.New("maxRunners must be greater than or equal to 0")
+	}
+	if c.MaxRunners > 0 && c.MinRunners > c.MaxRunners {
+		return errors.New("minRunners must be less than or equal to maxRunners")
+	}
+	return nil
+}
+
+// The Listener's role is to manage all interactions with the actions service.
+// It receives messages and processes them using the given handler.
+type Listener struct {
+	// configured fields
+	scaleSetID int               // The ID of the scale set associated with the listener.
+	client     Client            // The client used to interact with the scale set.
+	metrics    metrics.Publisher // The publisher used to publish metrics.
+
+	// internal fields
+	logger   logr.Logger // The logger used for logging.
+	hostname string      // The hostname of the listener.
+
+	// updated fields
+	lastMessageID int64                          // The ID of the last processed message.
+	maxCapacity   int                            // The maximum number of runners that can be created.
+	session       *actions.RunnerScaleSetSession // The session for managing the runner scale set.
+}
+
+func New(config Config) (*Listener, error) {
+	if err := config.Validate(); err != nil {
+		return nil, fmt.Errorf("invalid config: %w", err)
+	}
+
+	listener := &Listener{
+		scaleSetID:  config.ScaleSetID,
+		client:      config.Client,
+		logger:      config.Logger,
+		metrics:     metrics.Discard,
+		maxCapacity: config.MaxRunners,
+	}
+
+	if config.Metrics != nil {
+		listener.metrics = config.Metrics
+	}
+
+	listener.metrics.PublishStatic(config.MinRunners, config.MaxRunners)
+
+	hostname, err := os.Hostname()
+	if err != nil {
+		hostname = uuid.NewString()
+		listener.logger.Info("Failed to get hostname, fallback to uuid", "uuid", hostname, "error", err)
+	}
+	listener.hostname = hostname
+
+	return listener, nil
+}
+
+//go:generate mockery --name Handler --output ./mocks --outpkg mocks --case underscore
+type Handler interface {
+	HandleJobStarted(ctx context.Context, jobInfo *actions.JobStarted) error
+	HandleDesiredRunnerCount(ctx context.Context, count, jobsCompleted int) (int, error)
+}
+
+// Listen listens for incoming messages and handles them using the provided handler.
+// It continuously listens for messages until the context is cancelled.
+// The initial message contains the current statistics and acquirable jobs, if any.
+// The handler is responsible for handling the initial message and subsequent messages.
+// If an error occurs during any step, Listen returns an error.
+func (l *Listener) Listen(ctx context.Context, handler Handler) error {
+	if err := l.createSession(ctx); err != nil {
+		return fmt.Errorf("createSession failed: %w", err)
+	}
+
+	defer func() {
+		if err := l.deleteMessageSession(); err != nil {
+			l.logger.Error(err, "failed to delete message session")
+		}
+	}()
+
+	initialMessage := &actions.RunnerScaleSetMessage{
+		MessageId:   0,
+		MessageType: "RunnerScaleSetJobMessages",
+		Statistics:  l.session.Statistics,
+		Body:        "",
+	}
+
+	if l.session.Statistics == nil {
+		return fmt.Errorf("session statistics is nil")
+	}
+	l.metrics.PublishStatistics(initialMessage.Statistics)
+
+	desiredRunners, err := handler.HandleDesiredRunnerCount(ctx, initialMessage.Statistics.TotalAssignedJobs, 0)
+	if err != nil {
+		return fmt.Errorf("handling initial message failed: %w", err)
+	}
+	l.metrics.PublishDesiredRunners(desiredRunners)
+
+	for {
+		select {
+		case <-ctx.Done():
+			return ctx.Err()
+		default:
+		}
+
+		msg, err := l.getMessage(ctx)
+		if err != nil {
+			return fmt.Errorf("failed to get message: %w", err)
+		}
+
+		if msg == nil {
+			_, err := handler.HandleDesiredRunnerCount(ctx, 0, 0)
+			if err != nil {
+				return fmt.Errorf("handling nil message failed: %w", err)
+			}
+
+			continue
+		}
+
+		// Remove cancellation from the context to avoid cancelling the message handling.
+		if err := l.handleMessage(context.WithoutCancel(ctx), handler, msg); err != nil {
+			return fmt.Errorf("failed to handle message: %w", err)
+		}
+	}
+}
+
+func (l *Listener) handleMessage(ctx context.Context, handler Handler, msg *actions.RunnerScaleSetMessage) error {
+	parsedMsg, err := l.parseMessage(ctx, msg)
+	if err != nil {
+		return fmt.Errorf("failed to parse message: %w", err)
+	}
+	l.metrics.PublishStatistics(parsedMsg.statistics)
+
+	if len(parsedMsg.jobsAvailable) > 0 {
+		acquiredJobIDs, err := l.acquireAvailableJobs(ctx, parsedMsg.jobsAvailable)
+		if err != nil {
+			return fmt.Errorf("failed to acquire jobs: %w", err)
+		}
+
+		l.logger.Info("Jobs are acquired", "count", len(acquiredJobIDs), "requestIds", fmt.Sprint(acquiredJobIDs))
+	}
+
+	for _, jobCompleted := range parsedMsg.jobsCompleted {
+		l.metrics.PublishJobCompleted(jobCompleted)
+	}
+
+	l.lastMessageID = msg.MessageId
+
+	if err := l.deleteLastMessage(ctx); err != nil {
+		return fmt.Errorf("failed to delete message: %w", err)
+	}
+
+	for _, jobStarted := range parsedMsg.jobsStarted {
+		if err := handler.HandleJobStarted(ctx, jobStarted); err != nil {
+			return fmt.Errorf("failed to handle job started: %w", err)
+		}
+		l.metrics.PublishJobStarted(jobStarted)
+	}
+
+	desiredRunners, err := handler.HandleDesiredRunnerCount(ctx, parsedMsg.statistics.TotalAssignedJobs, len(parsedMsg.jobsCompleted))
+	if err != nil {
+		return fmt.Errorf("failed to handle desired runner count: %w", err)
+	}
+	l.metrics.PublishDesiredRunners(desiredRunners)
+	return nil
+}
+
+func (l *Listener) createSession(ctx context.Context) error {
+	var session *actions.RunnerScaleSetSession
+	var retries int
+
+	for {
+		var err error
+		session, err = l.client.CreateMessageSession(ctx, l.scaleSetID, l.hostname)
+		if err == nil {
+			break
+		}
+
+		clientErr := &actions.HttpClientSideError{}
+		if !errors.As(err, &clientErr) {
+			return fmt.Errorf("failed to create session: %w", err)
+		}
+
+		if clientErr.Code != http.StatusConflict {
+			return fmt.Errorf("failed to create session: %w", err)
+		}
+
+		retries++
+		if retries >= sessionCreationMaxRetries {
+			return fmt.Errorf("failed to create session after %d retries: %w", retries, err)
+		}
+
+		l.logger.Info("Unable to create message session. Will try again in 30 seconds", "error", err.Error())
+
+		select {
+		case <-ctx.Done():
+			return fmt.Errorf("context cancelled: %w", ctx.Err())
+		case <-time.After(30 * time.Second):
+		}
+	}
+
+	statistics, err := json.Marshal(session.Statistics)
+	if err != nil {
+		return fmt.Errorf("failed to marshal statistics: %w", err)
+	}
+	l.logger.Info("Current runner scale set statistics.", "statistics", string(statistics))
+
+	l.session = session
+
+	return nil
+}
+
+func (l *Listener) getMessage(ctx context.Context) (*actions.RunnerScaleSetMessage, error) {
+	l.logger.Info("Getting next message", "lastMessageID", l.lastMessageID)
+	msg, err := l.client.GetMessage(ctx, l.session.MessageQueueUrl, l.session.MessageQueueAccessToken, l.lastMessageID, l.maxCapacity)
+	if err == nil { // if NO error
+		return msg, nil
+	}
+
+	expiredError := &actions.MessageQueueTokenExpiredError{}
+	if !errors.As(err, &expiredError) {
+		return nil, fmt.Errorf("failed to get next message: %w", err)
+	}
+
+	if err := l.refreshSession(ctx); err != nil {
+		return nil, err
+	}
+
+	l.logger.Info("Getting next message", "lastMessageID", l.lastMessageID)
+
+	msg, err = l.client.GetMessage(ctx, l.session.MessageQueueUrl, l.session.MessageQueueAccessToken, l.lastMessageID, l.maxCapacity)
+	if err != nil { // if NO error
+		return nil, fmt.Errorf("failed to get next message after message session refresh: %w", err)
+	}
+
+	return msg, nil
+}
+
+func (l *Listener) deleteLastMessage(ctx context.Context) error {
+	l.logger.Info("Deleting last message", "lastMessageID", l.lastMessageID)
+	if err := l.client.DeleteMessage(ctx, l.session.MessageQueueUrl, l.session.MessageQueueAccessToken, l.lastMessageID); err != nil {
+		return fmt.Errorf("failed to delete message: %w", err)
+	}
+
+	return nil
+}
+
+type parsedMessage struct {
+	statistics    *actions.RunnerScaleSetStatistic
+	jobsStarted   []*actions.JobStarted
+	jobsAvailable []*actions.JobAvailable
+	jobsCompleted []*actions.JobCompleted
+}
+
+func (l *Listener) parseMessage(ctx context.Context, msg *actions.RunnerScaleSetMessage) (*parsedMessage, error) {
+	if msg.MessageType != "RunnerScaleSetJobMessages" {
+		l.logger.Info("Skipping message", "messageType", msg.MessageType)
+		return nil, fmt.Errorf("invalid message type: %s", msg.MessageType)
+	}
+
+	l.logger.Info("Processing message", "messageId", msg.MessageId, "messageType", msg.MessageType)
+	if msg.Statistics == nil {
+		return nil, fmt.Errorf("invalid message: statistics is nil")
+	}
+
+	l.logger.Info("New runner scale set statistics.", "statistics", msg.Statistics)
+
+	var batchedMessages []json.RawMessage
+	if len(msg.Body) > 0 {
+		if err := json.Unmarshal([]byte(msg.Body), &batchedMessages); err != nil {
+			return nil, fmt.Errorf("failed to unmarshal batched messages: %w", err)
+		}
+	}
+
+	parsedMsg := &parsedMessage{
+		statistics: msg.Statistics,
+	}
+
+	for _, msg := range batchedMessages {
+		var messageType actions.JobMessageType
+		if err := json.Unmarshal(msg, &messageType); err != nil {
+			return nil, fmt.Errorf("failed to decode job message type: %w", err)
+		}
+
+		switch messageType.MessageType {
+		case messageTypeJobAvailable:
+			var jobAvailable actions.JobAvailable
+			if err := json.Unmarshal(msg, &jobAvailable); err != nil {
+				return nil, fmt.Errorf("failed to decode job available: %w", err)
+			}
+
+			l.logger.Info("Job available message received", "jobId", jobAvailable.RunnerRequestId)
+			parsedMsg.jobsAvailable = append(parsedMsg.jobsAvailable, &jobAvailable)
+
+		case messageTypeJobAssigned:
+			var jobAssigned actions.JobAssigned
+			if err := json.Unmarshal(msg, &jobAssigned); err != nil {
+				return nil, fmt.Errorf("failed to decode job assigned: %w", err)
+			}
+
+			l.logger.Info("Job assigned message received", "jobId", jobAssigned.RunnerRequestId)
+
+		case messageTypeJobStarted:
+			var jobStarted actions.JobStarted
+			if err := json.Unmarshal(msg, &jobStarted); err != nil {
+				return nil, fmt.Errorf("could not decode job started message. %w", err)
+			}
+			l.logger.Info("Job started message received.", "RequestId", jobStarted.RunnerRequestId, "RunnerId", jobStarted.RunnerId)
+			parsedMsg.jobsStarted = append(parsedMsg.jobsStarted, &jobStarted)
+
+		case messageTypeJobCompleted:
+			var jobCompleted actions.JobCompleted
+			if err := json.Unmarshal(msg, &jobCompleted); err != nil {
+				return nil, fmt.Errorf("failed to decode job completed: %w", err)
+			}
+
+			l.logger.Info("Job completed message received.", "RequestId", jobCompleted.RunnerRequestId, "Result", jobCompleted.Result, "RunnerId", jobCompleted.RunnerId, "RunnerName", jobCompleted.RunnerName)
+			parsedMsg.jobsCompleted = append(parsedMsg.jobsCompleted, &jobCompleted)
+
+		default:
+			l.logger.Info("unknown job message type.", "messageType", messageType.MessageType)
+		}
+	}
+
+	return parsedMsg, nil
+}
+
+func (l *Listener) acquireAvailableJobs(ctx context.Context, jobsAvailable []*actions.JobAvailable) ([]int64, error) {
+	ids := make([]int64, 0, len(jobsAvailable))
+	for _, job := range jobsAvailable {
+		ids = append(ids, job.RunnerRequestId)
+	}
+
+	l.logger.Info("Acquiring jobs", "count", len(ids), "requestIds", fmt.Sprint(ids))
+
+	idsAcquired, err := l.client.AcquireJobs(ctx, l.scaleSetID, l.session.MessageQueueAccessToken, ids)
+	if err == nil { // if NO errors
+		return idsAcquired, nil
+	}
+
+	expiredError := &actions.MessageQueueTokenExpiredError{}
+	if !errors.As(err, &expiredError) {
+		return nil, fmt.Errorf("failed to acquire jobs: %w", err)
+	}
+
+	if err := l.refreshSession(ctx); err != nil {
+		return nil, err
+	}
+
+	idsAcquired, err = l.client.AcquireJobs(ctx, l.scaleSetID, l.session.MessageQueueAccessToken, ids)
+	if err != nil {
+		return nil, fmt.Errorf("failed to acquire jobs after session refresh: %w", err)
+	}
+
+	return idsAcquired, nil
+}
+
+func (l *Listener) refreshSession(ctx context.Context) error {
+	l.logger.Info("Message queue token is expired during GetNextMessage, refreshing...")
+	session, err := l.client.RefreshMessageSession(ctx, l.session.RunnerScaleSet.Id, l.session.SessionId)
+	if err != nil {
+		return fmt.Errorf("refresh message session failed. %w", err)
+	}
+
+	l.session = session
+	return nil
+}
+
+func (l *Listener) deleteMessageSession() error {
+	ctx, cancel := context.WithTimeout(context.Background(), 30*time.Second)
+	defer cancel()
+
+	l.logger.Info("Deleting message session")
+
+	if err := l.client.DeleteMessageSession(ctx, l.session.RunnerScaleSet.Id, l.session.SessionId); err != nil {
+		return fmt.Errorf("failed to delete message session: %w", err)
+	}
+
+	return nil
+}

cmd/ghalistener/listener/listener_test.go

@@ -0,0 +1,883 @@
+package listener
+
+import (
+	"context"
+	"encoding/json"
+	"errors"
+	"net/http"
+	"testing"
+	"time"
+
+	listenermocks "github.com/actions/actions-runner-controller/cmd/ghalistener/listener/mocks"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/metrics"
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/google/uuid"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/mock"
+	"github.com/stretchr/testify/require"
+)
+
+func TestNew(t *testing.T) {
+	t.Parallel()
+	t.Run("InvalidConfig", func(t *testing.T) {
+		t.Parallel()
+		var config Config
+		_, err := New(config)
+		assert.NotNil(t, err)
+	})
+
+	t.Run("ValidConfig", func(t *testing.T) {
+		t.Parallel()
+		config := Config{
+			Client:     listenermocks.NewClient(t),
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+		l, err := New(config)
+		assert.Nil(t, err)
+		assert.NotNil(t, l)
+	})
+}
+
+func TestListener_createSession(t *testing.T) {
+	t.Parallel()
+	t.Run("FailOnce", func(t *testing.T) {
+		t.Parallel()
+		ctx := context.Background()
+
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+		client.On("CreateMessageSession", ctx, mock.Anything, mock.Anything).Return(nil, assert.AnError).Once()
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		err = l.createSession(ctx)
+		assert.NotNil(t, err)
+	})
+
+	t.Run("FailContext", func(t *testing.T) {
+		t.Parallel()
+		ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
+		defer cancel()
+
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+		client.On("CreateMessageSession", ctx, mock.Anything, mock.Anything).Return(nil,
+			&actions.HttpClientSideError{Code: http.StatusConflict}).Once()
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		err = l.createSession(ctx)
+		assert.True(t, errors.Is(err, context.DeadlineExceeded))
+	})
+
+	t.Run("SetsSession", func(t *testing.T) {
+		t.Parallel()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              nil,
+		}
+		client.On("CreateMessageSession", mock.Anything, mock.Anything, mock.Anything).Return(session, nil).Once()
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		err = l.createSession(context.Background())
+		assert.Nil(t, err)
+		assert.Equal(t, session, l.session)
+	})
+}
+
+func TestListener_getMessage(t *testing.T) {
+	t.Parallel()
+
+	t.Run("ReceivesMessage", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+			MaxRunners: 10,
+		}
+
+		client := listenermocks.NewClient(t)
+		want := &actions.RunnerScaleSetMessage{
+			MessageId: 1,
+		}
+		client.On("GetMessage", ctx, mock.Anything, mock.Anything, mock.Anything, 10).Return(want, nil).Once()
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+		l.session = &actions.RunnerScaleSetSession{}
+
+		got, err := l.getMessage(ctx)
+		assert.Nil(t, err)
+		assert.Equal(t, want, got)
+	})
+
+	t.Run("NotExpiredError", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+			MaxRunners: 10,
+		}
+
+		client := listenermocks.NewClient(t)
+		client.On("GetMessage", ctx, mock.Anything, mock.Anything, mock.Anything, 10).Return(nil, &actions.HttpClientSideError{Code: http.StatusNotFound}).Once()
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		l.session = &actions.RunnerScaleSetSession{}
+
+		_, err = l.getMessage(ctx)
+		assert.NotNil(t, err)
+	})
+
+	t.Run("RefreshAndSucceeds", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+			MaxRunners: 10,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              nil,
+		}
+		client.On("RefreshMessageSession", ctx, mock.Anything, mock.Anything).Return(session, nil).Once()
+
+		client.On("GetMessage", ctx, mock.Anything, mock.Anything, mock.Anything, 10).Return(nil, &actions.MessageQueueTokenExpiredError{}).Once()
+
+		want := &actions.RunnerScaleSetMessage{
+			MessageId: 1,
+		}
+		client.On("GetMessage", ctx, mock.Anything, mock.Anything, mock.Anything, 10).Return(want, nil).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		l.session = &actions.RunnerScaleSetSession{
+			SessionId:      &uuid,
+			RunnerScaleSet: &actions.RunnerScaleSet{},
+		}
+
+		got, err := l.getMessage(ctx)
+		assert.Nil(t, err)
+		assert.Equal(t, want, got)
+	})
+
+	t.Run("RefreshAndFails", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+			MaxRunners: 10,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              nil,
+		}
+		client.On("RefreshMessageSession", ctx, mock.Anything, mock.Anything).Return(session, nil).Once()
+
+		client.On("GetMessage", ctx, mock.Anything, mock.Anything, mock.Anything, 10).Return(nil, &actions.MessageQueueTokenExpiredError{}).Twice()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		l.session = &actions.RunnerScaleSetSession{
+			SessionId:      &uuid,
+			RunnerScaleSet: &actions.RunnerScaleSet{},
+		}
+
+		got, err := l.getMessage(ctx)
+		assert.NotNil(t, err)
+		assert.Nil(t, got)
+	})
+}
+
+func TestListener_refreshSession(t *testing.T) {
+	t.Parallel()
+
+	t.Run("SuccessfullyRefreshes", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		newUUID := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &newUUID,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              nil,
+		}
+		client.On("RefreshMessageSession", ctx, mock.Anything, mock.Anything).Return(session, nil).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		oldUUID := uuid.New()
+		l.session = &actions.RunnerScaleSetSession{
+			SessionId:      &oldUUID,
+			RunnerScaleSet: &actions.RunnerScaleSet{},
+		}
+
+		err = l.refreshSession(ctx)
+		assert.Nil(t, err)
+		assert.Equal(t, session, l.session)
+	})
+
+	t.Run("FailsToRefresh", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		client.On("RefreshMessageSession", ctx, mock.Anything, mock.Anything).Return(nil, errors.New("error")).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		oldUUID := uuid.New()
+		oldSession := &actions.RunnerScaleSetSession{
+			SessionId:      &oldUUID,
+			RunnerScaleSet: &actions.RunnerScaleSet{},
+		}
+		l.session = oldSession
+
+		err = l.refreshSession(ctx)
+		assert.NotNil(t, err)
+		assert.Equal(t, oldSession, l.session)
+	})
+}
+
+func TestListener_deleteLastMessage(t *testing.T) {
+	t.Parallel()
+
+	t.Run("SuccessfullyDeletes", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		client.On("DeleteMessage", ctx, mock.Anything, mock.Anything, mock.MatchedBy(func(lastMessageID any) bool {
+			return lastMessageID.(int64) == int64(5)
+		})).Return(nil).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		l.session = &actions.RunnerScaleSetSession{}
+		l.lastMessageID = 5
+
+		err = l.deleteLastMessage(ctx)
+		assert.Nil(t, err)
+	})
+
+	t.Run("FailsToDelete", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		client.On("DeleteMessage", ctx, mock.Anything, mock.Anything, mock.Anything).Return(errors.New("error")).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		l.session = &actions.RunnerScaleSetSession{}
+		l.lastMessageID = 5
+
+		err = l.deleteLastMessage(ctx)
+		assert.NotNil(t, err)
+	})
+}
+
+func TestListener_Listen(t *testing.T) {
+	t.Parallel()
+
+	t.Run("CreateSessionFails", func(t *testing.T) {
+		t.Parallel()
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+		client.On("CreateMessageSession", ctx, mock.Anything, mock.Anything).Return(nil, assert.AnError).Once()
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		err = l.Listen(ctx, nil)
+		assert.NotNil(t, err)
+	})
+
+	t.Run("CallHandleRegardlessOfInitialMessage", func(t *testing.T) {
+		t.Parallel()
+		ctx, cancel := context.WithCancel(context.Background())
+
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              &actions.RunnerScaleSetStatistic{},
+		}
+		client.On("CreateMessageSession", ctx, mock.Anything, mock.Anything).Return(session, nil).Once()
+		client.On("DeleteMessageSession", mock.Anything, session.RunnerScaleSet.Id, session.SessionId).Return(nil).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		var called bool
+		handler := listenermocks.NewHandler(t)
+		handler.On("HandleDesiredRunnerCount", mock.Anything, mock.Anything, 0).
+			Return(0, nil).
+			Run(
+				func(mock.Arguments) {
+					called = true
+					cancel()
+				},
+			).
+			Once()
+
+		err = l.Listen(ctx, handler)
+		assert.True(t, errors.Is(err, context.Canceled))
+		assert.True(t, called)
+	})
+
+	t.Run("CancelContextAfterGetMessage", func(t *testing.T) {
+		t.Parallel()
+
+		ctx, cancel := context.WithCancel(context.Background())
+
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+			MaxRunners: 10,
+		}
+
+		client := listenermocks.NewClient(t)
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              &actions.RunnerScaleSetStatistic{},
+		}
+		client.On("CreateMessageSession", ctx, mock.Anything, mock.Anything).Return(session, nil).Once()
+		client.On("DeleteMessageSession", mock.Anything, session.RunnerScaleSet.Id, session.SessionId).Return(nil).Once()
+
+		msg := &actions.RunnerScaleSetMessage{
+			MessageId:   1,
+			MessageType: "RunnerScaleSetJobMessages",
+			Statistics:  &actions.RunnerScaleSetStatistic{},
+		}
+		client.On("GetMessage", ctx, mock.Anything, mock.Anything, mock.Anything, 10).
+			Return(msg, nil).
+			Run(
+				func(mock.Arguments) {
+					cancel()
+				},
+			).
+			Once()
+
+		// Ensure delete message is called without cancel
+		client.On("DeleteMessage", context.WithoutCancel(ctx), mock.Anything, mock.Anything, mock.Anything).Return(nil).Once()
+
+		config.Client = client
+
+		handler := listenermocks.NewHandler(t)
+		handler.On("HandleDesiredRunnerCount", mock.Anything, mock.Anything, 0).
+			Return(0, nil).
+			Once()
+
+		handler.On("HandleDesiredRunnerCount", mock.Anything, mock.Anything, 0).
+			Return(0, nil).
+			Once()
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		err = l.Listen(ctx, handler)
+		assert.ErrorIs(t, context.Canceled, err)
+	})
+}
+
+func TestListener_acquireAvailableJobs(t *testing.T) {
+	t.Parallel()
+
+	t.Run("FailingToAcquireJobs", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		client.On("AcquireJobs", ctx, mock.Anything, mock.Anything, mock.Anything).Return(nil, assert.AnError).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		uuid := uuid.New()
+		l.session = &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              &actions.RunnerScaleSetStatistic{},
+		}
+
+		availableJobs := []*actions.JobAvailable{
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 1,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 2,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 3,
+				},
+			},
+		}
+		_, err = l.acquireAvailableJobs(ctx, availableJobs)
+		assert.Error(t, err)
+	})
+
+	t.Run("SuccessfullyAcquiresJobsOnFirstRun", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		jobIDs := []int64{1, 2, 3}
+
+		client.On("AcquireJobs", ctx, mock.Anything, mock.Anything, mock.Anything).Return(jobIDs, nil).Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		uuid := uuid.New()
+		l.session = &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              &actions.RunnerScaleSetStatistic{},
+		}
+
+		availableJobs := []*actions.JobAvailable{
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 1,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 2,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 3,
+				},
+			},
+		}
+		acquiredJobIDs, err := l.acquireAvailableJobs(ctx, availableJobs)
+		assert.NoError(t, err)
+		assert.Equal(t, []int64{1, 2, 3}, acquiredJobIDs)
+	})
+
+	t.Run("RefreshAndSucceeds", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              nil,
+		}
+		client.On("RefreshMessageSession", ctx, mock.Anything, mock.Anything).Return(session, nil).Once()
+
+		// Second call to AcquireJobs will succeed
+		want := []int64{1, 2, 3}
+		availableJobs := []*actions.JobAvailable{
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 1,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 2,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 3,
+				},
+			},
+		}
+
+		// First call to AcquireJobs will fail with a token expired error
+		client.On("AcquireJobs", ctx, mock.Anything, mock.Anything, mock.Anything).
+			Run(func(args mock.Arguments) {
+				ids := args.Get(3).([]int64)
+				assert.Equal(t, want, ids)
+			}).
+			Return(nil, &actions.MessageQueueTokenExpiredError{}).
+			Once()
+
+		// Second call should succeed
+		client.On("AcquireJobs", ctx, mock.Anything, mock.Anything, mock.Anything).
+			Run(func(args mock.Arguments) {
+				ids := args.Get(3).([]int64)
+				assert.Equal(t, want, ids)
+			}).
+			Return(want, nil).
+			Once()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		l.session = &actions.RunnerScaleSetSession{
+			SessionId:      &uuid,
+			RunnerScaleSet: &actions.RunnerScaleSet{},
+		}
+
+		got, err := l.acquireAvailableJobs(ctx, availableJobs)
+		assert.Nil(t, err)
+		assert.Equal(t, want, got)
+	})
+
+	t.Run("RefreshAndFails", func(t *testing.T) {
+		t.Parallel()
+
+		ctx := context.Background()
+		config := Config{
+			ScaleSetID: 1,
+			Metrics:    metrics.Discard,
+		}
+
+		client := listenermocks.NewClient(t)
+
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              nil,
+		}
+		client.On("RefreshMessageSession", ctx, mock.Anything, mock.Anything).Return(session, nil).Once()
+
+		client.On("AcquireJobs", ctx, mock.Anything, mock.Anything, mock.Anything).Return(nil, &actions.MessageQueueTokenExpiredError{}).Twice()
+
+		config.Client = client
+
+		l, err := New(config)
+		require.Nil(t, err)
+
+		l.session = &actions.RunnerScaleSetSession{
+			SessionId:      &uuid,
+			RunnerScaleSet: &actions.RunnerScaleSet{},
+		}
+
+		availableJobs := []*actions.JobAvailable{
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 1,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 2,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					RunnerRequestId: 3,
+				},
+			},
+		}
+
+		got, err := l.acquireAvailableJobs(ctx, availableJobs)
+		assert.NotNil(t, err)
+		assert.Nil(t, got)
+	})
+}
+
+func TestListener_parseMessage(t *testing.T) {
+	t.Run("FailOnEmptyStatistics", func(t *testing.T) {
+		msg := &actions.RunnerScaleSetMessage{
+			MessageId:   1,
+			MessageType: "RunnerScaleSetJobMessages",
+			Statistics:  nil,
+		}
+
+		l := &Listener{}
+		parsedMsg, err := l.parseMessage(context.Background(), msg)
+		assert.Error(t, err)
+		assert.Nil(t, parsedMsg)
+	})
+
+	t.Run("FailOnIncorrectMessageType", func(t *testing.T) {
+		msg := &actions.RunnerScaleSetMessage{
+			MessageId:   1,
+			MessageType: "RunnerMessages", // arbitrary message type
+			Statistics:  &actions.RunnerScaleSetStatistic{},
+		}
+
+		l := &Listener{}
+		parsedMsg, err := l.parseMessage(context.Background(), msg)
+		assert.Error(t, err)
+		assert.Nil(t, parsedMsg)
+	})
+
+	t.Run("ParseAll", func(t *testing.T) {
+		msg := &actions.RunnerScaleSetMessage{
+			MessageId:   1,
+			MessageType: "RunnerScaleSetJobMessages",
+			Body:        "",
+			Statistics: &actions.RunnerScaleSetStatistic{
+				TotalAvailableJobs:     1,
+				TotalAcquiredJobs:      2,
+				TotalAssignedJobs:      3,
+				TotalRunningJobs:       4,
+				TotalRegisteredRunners: 5,
+				TotalBusyRunners:       6,
+				TotalIdleRunners:       7,
+			},
+		}
+
+		var batchedMessages []any
+		jobsAvailable := []*actions.JobAvailable{
+			{
+				AcquireJobUrl: "https://github.com/example",
+				JobMessageBase: actions.JobMessageBase{
+					JobMessageType: actions.JobMessageType{
+						MessageType: messageTypeJobAvailable,
+					},
+					RunnerRequestId: 1,
+				},
+			},
+			{
+				AcquireJobUrl: "https://github.com/example",
+				JobMessageBase: actions.JobMessageBase{
+					JobMessageType: actions.JobMessageType{
+						MessageType: messageTypeJobAvailable,
+					},
+					RunnerRequestId: 2,
+				},
+			},
+		}
+		for _, msg := range jobsAvailable {
+			batchedMessages = append(batchedMessages, msg)
+		}
+
+		jobsAssigned := []*actions.JobAssigned{
+			{
+				JobMessageBase: actions.JobMessageBase{
+					JobMessageType: actions.JobMessageType{
+						MessageType: messageTypeJobAssigned,
+					},
+					RunnerRequestId: 3,
+				},
+			},
+			{
+				JobMessageBase: actions.JobMessageBase{
+					JobMessageType: actions.JobMessageType{
+						MessageType: messageTypeJobAssigned,
+					},
+					RunnerRequestId: 4,
+				},
+			},
+		}
+		for _, msg := range jobsAssigned {
+			batchedMessages = append(batchedMessages, msg)
+		}
+
+		jobsStarted := []*actions.JobStarted{
+			{
+				JobMessageBase: actions.JobMessageBase{
+					JobMessageType: actions.JobMessageType{
+						MessageType: messageTypeJobStarted,
+					},
+					RunnerRequestId: 5,
+				},
+				RunnerId:   2,
+				RunnerName: "runner2",
+			},
+		}
+		for _, msg := range jobsStarted {
+			batchedMessages = append(batchedMessages, msg)
+		}
+
+		jobsCompleted := []*actions.JobCompleted{
+			{
+				JobMessageBase: actions.JobMessageBase{
+					JobMessageType: actions.JobMessageType{
+						MessageType: messageTypeJobCompleted,
+					},
+					RunnerRequestId: 6,
+				},
+				Result:     "success",
+				RunnerId:   1,
+				RunnerName: "runner1",
+			},
+		}
+		for _, msg := range jobsCompleted {
+			batchedMessages = append(batchedMessages, msg)
+		}
+
+		b, err := json.Marshal(batchedMessages)
+		require.NoError(t, err)
+
+		msg.Body = string(b)
+
+		l := &Listener{}
+		parsedMsg, err := l.parseMessage(context.Background(), msg)
+		require.NoError(t, err)
+
+		assert.Equal(t, msg.Statistics, parsedMsg.statistics)
+		assert.Equal(t, jobsAvailable, parsedMsg.jobsAvailable)
+		assert.Equal(t, jobsStarted, parsedMsg.jobsStarted)
+		assert.Equal(t, jobsCompleted, parsedMsg.jobsCompleted)
+	})
+}

cmd/ghalistener/listener/metrics_test.go

@@ -0,0 +1,205 @@
+package listener
+
+import (
+	"context"
+	"encoding/json"
+	"testing"
+
+	listenermocks "github.com/actions/actions-runner-controller/cmd/ghalistener/listener/mocks"
+	metricsmocks "github.com/actions/actions-runner-controller/cmd/ghalistener/metrics/mocks"
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/google/uuid"
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/mock"
+	"github.com/stretchr/testify/require"
+)
+
+func TestInitialMetrics(t *testing.T) {
+	t.Parallel()
+
+	t.Run("SetStaticMetrics", func(t *testing.T) {
+		t.Parallel()
+
+		metrics := metricsmocks.NewPublisher(t)
+
+		minRunners := 5
+		maxRunners := 10
+		metrics.On("PublishStatic", minRunners, maxRunners).Once()
+
+		config := Config{
+			Client:     listenermocks.NewClient(t),
+			ScaleSetID: 1,
+			Metrics:    metrics,
+			MinRunners: minRunners,
+			MaxRunners: maxRunners,
+		}
+		l, err := New(config)
+
+		assert.Nil(t, err)
+		assert.NotNil(t, l)
+	})
+
+	t.Run("InitialMessageStatistics", func(t *testing.T) {
+		t.Parallel()
+
+		ctx, cancel := context.WithCancel(context.Background())
+
+		sessionStatistics := &actions.RunnerScaleSetStatistic{
+			TotalAvailableJobs:     1,
+			TotalAcquiredJobs:      2,
+			TotalAssignedJobs:      3,
+			TotalRunningJobs:       4,
+			TotalRegisteredRunners: 5,
+			TotalBusyRunners:       6,
+			TotalIdleRunners:       7,
+		}
+
+		uuid := uuid.New()
+		session := &actions.RunnerScaleSetSession{
+			SessionId:               &uuid,
+			OwnerName:               "example",
+			RunnerScaleSet:          &actions.RunnerScaleSet{},
+			MessageQueueUrl:         "https://example.com",
+			MessageQueueAccessToken: "1234567890",
+			Statistics:              sessionStatistics,
+		}
+
+		metrics := metricsmocks.NewPublisher(t)
+		metrics.On("PublishStatic", mock.Anything, mock.Anything).Once()
+		metrics.On("PublishStatistics", sessionStatistics).Once()
+		metrics.On("PublishDesiredRunners", sessionStatistics.TotalAssignedJobs).
+			Run(
+				func(mock.Arguments) {
+					cancel()
+				},
+			).Once()
+
+		config := Config{
+			Client:     listenermocks.NewClient(t),
+			ScaleSetID: 1,
+			Metrics:    metrics,
+		}
+
+		client := listenermocks.NewClient(t)
+		client.On("CreateMessageSession", mock.Anything, mock.Anything, mock.Anything).Return(session, nil).Once()
+		client.On("DeleteMessageSession", mock.Anything, session.RunnerScaleSet.Id, session.SessionId).Return(nil).Once()
+		config.Client = client
+
+		handler := listenermocks.NewHandler(t)
+		handler.On("HandleDesiredRunnerCount", mock.Anything, sessionStatistics.TotalAssignedJobs, 0).
+			Return(sessionStatistics.TotalAssignedJobs, nil).
+			Once()
+
+		l, err := New(config)
+		assert.Nil(t, err)
+		assert.NotNil(t, l)
+
+		assert.ErrorIs(t, context.Canceled, l.Listen(ctx, handler))
+	})
+}
+
+func TestHandleMessageMetrics(t *testing.T) {
+	t.Parallel()
+
+	msg := &actions.RunnerScaleSetMessage{
+		MessageId:   1,
+		MessageType: "RunnerScaleSetJobMessages",
+		Body:        "",
+		Statistics: &actions.RunnerScaleSetStatistic{
+			TotalAvailableJobs:     1,
+			TotalAcquiredJobs:      2,
+			TotalAssignedJobs:      3,
+			TotalRunningJobs:       4,
+			TotalRegisteredRunners: 5,
+			TotalBusyRunners:       6,
+			TotalIdleRunners:       7,
+		},
+	}
+
+	var batchedMessages []any
+	jobsStarted := []*actions.JobStarted{
+		{
+			JobMessageBase: actions.JobMessageBase{
+				JobMessageType: actions.JobMessageType{
+					MessageType: messageTypeJobStarted,
+				},
+				RunnerRequestId: 8,
+			},
+			RunnerId:   3,
+			RunnerName: "runner3",
+		},
+	}
+	for _, msg := range jobsStarted {
+		batchedMessages = append(batchedMessages, msg)
+	}
+
+	jobsCompleted := []*actions.JobCompleted{
+		{
+			JobMessageBase: actions.JobMessageBase{
+				JobMessageType: actions.JobMessageType{
+					MessageType: messageTypeJobCompleted,
+				},
+				RunnerRequestId: 6,
+			},
+			Result:     "success",
+			RunnerId:   1,
+			RunnerName: "runner1",
+		},
+		{
+			JobMessageBase: actions.JobMessageBase{
+				JobMessageType: actions.JobMessageType{
+					MessageType: messageTypeJobCompleted,
+				},
+				RunnerRequestId: 7,
+			},
+			Result:     "success",
+			RunnerId:   2,
+			RunnerName: "runner2",
+		},
+	}
+	for _, msg := range jobsCompleted {
+		batchedMessages = append(batchedMessages, msg)
+	}
+
+	b, err := json.Marshal(batchedMessages)
+	require.NoError(t, err)
+
+	msg.Body = string(b)
+
+	desiredResult := 4
+
+	metrics := metricsmocks.NewPublisher(t)
+	metrics.On("PublishStatic", 0, 0).Once()
+	metrics.On("PublishStatistics", msg.Statistics).Once()
+	metrics.On("PublishJobCompleted", jobsCompleted[0]).Once()
+	metrics.On("PublishJobCompleted", jobsCompleted[1]).Once()
+	metrics.On("PublishJobStarted", jobsStarted[0]).Once()
+	metrics.On("PublishDesiredRunners", desiredResult).Once()
+
+	handler := listenermocks.NewHandler(t)
+	handler.On("HandleJobStarted", mock.Anything, jobsStarted[0]).Return(nil).Once()
+	handler.On("HandleDesiredRunnerCount", mock.Anything, mock.Anything, 2).Return(desiredResult, nil).Once()
+
+	client := listenermocks.NewClient(t)
+	client.On("DeleteMessage", mock.Anything, mock.Anything, mock.Anything, mock.Anything).Return(nil).Once()
+
+	config := Config{
+		Client:     listenermocks.NewClient(t),
+		ScaleSetID: 1,
+		Metrics:    metrics,
+	}
+
+	l, err := New(config)
+	require.NoError(t, err)
+	l.client = client
+	l.session = &actions.RunnerScaleSetSession{
+		OwnerName:               "",
+		RunnerScaleSet:          &actions.RunnerScaleSet{},
+		MessageQueueUrl:         "",
+		MessageQueueAccessToken: "",
+		Statistics:              &actions.RunnerScaleSetStatistic{},
+	}
+
+	err = l.handleMessage(context.Background(), handler, msg)
+	require.NoError(t, err)
+}

cmd/ghalistener/listener/mocks/client.go

@@ -0,0 +1,190 @@
+// Code generated by mockery v2.36.1. DO NOT EDIT.
+
+package mocks
+
+import (
+	context "context"
+
+	actions "github.com/actions/actions-runner-controller/github/actions"
+
+	mock "github.com/stretchr/testify/mock"
+
+	uuid "github.com/google/uuid"
+)
+
+// Client is an autogenerated mock type for the Client type
+type Client struct {
+	mock.Mock
+}
+
+// AcquireJobs provides a mock function with given fields: ctx, runnerScaleSetId, messageQueueAccessToken, requestIds
+func (_m *Client) AcquireJobs(ctx context.Context, runnerScaleSetId int, messageQueueAccessToken string, requestIds []int64) ([]int64, error) {
+	ret := _m.Called(ctx, runnerScaleSetId, messageQueueAccessToken, requestIds)
+
+	var r0 []int64
+	var r1 error
+	if rf, ok := ret.Get(0).(func(context.Context, int, string, []int64) ([]int64, error)); ok {
+		return rf(ctx, runnerScaleSetId, messageQueueAccessToken, requestIds)
+	}
+	if rf, ok := ret.Get(0).(func(context.Context, int, string, []int64) []int64); ok {
+		r0 = rf(ctx, runnerScaleSetId, messageQueueAccessToken, requestIds)
+	} else {
+		if ret.Get(0) != nil {
+			r0 = ret.Get(0).([]int64)
+		}
+	}
+
+	if rf, ok := ret.Get(1).(func(context.Context, int, string, []int64) error); ok {
+		r1 = rf(ctx, runnerScaleSetId, messageQueueAccessToken, requestIds)
+	} else {
+		r1 = ret.Error(1)
+	}
+
+	return r0, r1
+}
+
+// CreateMessageSession provides a mock function with given fields: ctx, runnerScaleSetId, owner
+func (_m *Client) CreateMessageSession(ctx context.Context, runnerScaleSetId int, owner string) (*actions.RunnerScaleSetSession, error) {
+	ret := _m.Called(ctx, runnerScaleSetId, owner)
+
+	var r0 *actions.RunnerScaleSetSession
+	var r1 error
+	if rf, ok := ret.Get(0).(func(context.Context, int, string) (*actions.RunnerScaleSetSession, error)); ok {
+		return rf(ctx, runnerScaleSetId, owner)
+	}
+	if rf, ok := ret.Get(0).(func(context.Context, int, string) *actions.RunnerScaleSetSession); ok {
+		r0 = rf(ctx, runnerScaleSetId, owner)
+	} else {
+		if ret.Get(0) != nil {
+			r0 = ret.Get(0).(*actions.RunnerScaleSetSession)
+		}
+	}
+
+	if rf, ok := ret.Get(1).(func(context.Context, int, string) error); ok {
+		r1 = rf(ctx, runnerScaleSetId, owner)
+	} else {
+		r1 = ret.Error(1)
+	}
+
+	return r0, r1
+}
+
+// DeleteMessage provides a mock function with given fields: ctx, messageQueueUrl, messageQueueAccessToken, messageId
+func (_m *Client) DeleteMessage(ctx context.Context, messageQueueUrl string, messageQueueAccessToken string, messageId int64) error {
+	ret := _m.Called(ctx, messageQueueUrl, messageQueueAccessToken, messageId)
+
+	var r0 error
+	if rf, ok := ret.Get(0).(func(context.Context, string, string, int64) error); ok {
+		r0 = rf(ctx, messageQueueUrl, messageQueueAccessToken, messageId)
+	} else {
+		r0 = ret.Error(0)
+	}
+
+	return r0
+}
+
+// DeleteMessageSession provides a mock function with given fields: ctx, runnerScaleSetId, sessionId
+func (_m *Client) DeleteMessageSession(ctx context.Context, runnerScaleSetId int, sessionId *uuid.UUID) error {
+	ret := _m.Called(ctx, runnerScaleSetId, sessionId)
+
+	var r0 error
+	if rf, ok := ret.Get(0).(func(context.Context, int, *uuid.UUID) error); ok {
+		r0 = rf(ctx, runnerScaleSetId, sessionId)
+	} else {
+		r0 = ret.Error(0)
+	}
+
+	return r0
+}
+
+// GetAcquirableJobs provides a mock function with given fields: ctx, runnerScaleSetId
+func (_m *Client) GetAcquirableJobs(ctx context.Context, runnerScaleSetId int) (*actions.AcquirableJobList, error) {
+	ret := _m.Called(ctx, runnerScaleSetId)
+
+	var r0 *actions.AcquirableJobList
+	var r1 error
+	if rf, ok := ret.Get(0).(func(context.Context, int) (*actions.AcquirableJobList, error)); ok {
+		return rf(ctx, runnerScaleSetId)
+	}
+	if rf, ok := ret.Get(0).(func(context.Context, int) *actions.AcquirableJobList); ok {
+		r0 = rf(ctx, runnerScaleSetId)
+	} else {
+		if ret.Get(0) != nil {
+			r0 = ret.Get(0).(*actions.AcquirableJobList)
+		}
+	}
+
+	if rf, ok := ret.Get(1).(func(context.Context, int) error); ok {
+		r1 = rf(ctx, runnerScaleSetId)
+	} else {
+		r1 = ret.Error(1)
+	}
+
+	return r0, r1
+}
+
+// GetMessage provides a mock function with given fields: ctx, messageQueueUrl, messageQueueAccessToken, lastMessageId, maxCapacity
+func (_m *Client) GetMessage(ctx context.Context, messageQueueUrl string, messageQueueAccessToken string, lastMessageId int64, maxCapacity int) (*actions.RunnerScaleSetMessage, error) {
+	ret := _m.Called(ctx, messageQueueUrl, messageQueueAccessToken, lastMessageId, maxCapacity)
+
+	var r0 *actions.RunnerScaleSetMessage
+	var r1 error
+	if rf, ok := ret.Get(0).(func(context.Context, string, string, int64, int) (*actions.RunnerScaleSetMessage, error)); ok {
+		return rf(ctx, messageQueueUrl, messageQueueAccessToken, lastMessageId, maxCapacity)
+	}
+	if rf, ok := ret.Get(0).(func(context.Context, string, string, int64, int) *actions.RunnerScaleSetMessage); ok {
+		r0 = rf(ctx, messageQueueUrl, messageQueueAccessToken, lastMessageId, maxCapacity)
+	} else {
+		if ret.Get(0) != nil {
+			r0 = ret.Get(0).(*actions.RunnerScaleSetMessage)
+		}
+	}
+
+	if rf, ok := ret.Get(1).(func(context.Context, string, string, int64, int) error); ok {
+		r1 = rf(ctx, messageQueueUrl, messageQueueAccessToken, lastMessageId, maxCapacity)
+	} else {
+		r1 = ret.Error(1)
+	}
+
+	return r0, r1
+}
+
+// RefreshMessageSession provides a mock function with given fields: ctx, runnerScaleSetId, sessionId
+func (_m *Client) RefreshMessageSession(ctx context.Context, runnerScaleSetId int, sessionId *uuid.UUID) (*actions.RunnerScaleSetSession, error) {
+	ret := _m.Called(ctx, runnerScaleSetId, sessionId)
+
+	var r0 *actions.RunnerScaleSetSession
+	var r1 error
+	if rf, ok := ret.Get(0).(func(context.Context, int, *uuid.UUID) (*actions.RunnerScaleSetSession, error)); ok {
+		return rf(ctx, runnerScaleSetId, sessionId)
+	}
+	if rf, ok := ret.Get(0).(func(context.Context, int, *uuid.UUID) *actions.RunnerScaleSetSession); ok {
+		r0 = rf(ctx, runnerScaleSetId, sessionId)
+	} else {
+		if ret.Get(0) != nil {
+			r0 = ret.Get(0).(*actions.RunnerScaleSetSession)
+		}
+	}
+
+	if rf, ok := ret.Get(1).(func(context.Context, int, *uuid.UUID) error); ok {
+		r1 = rf(ctx, runnerScaleSetId, sessionId)
+	} else {
+		r1 = ret.Error(1)
+	}
+
+	return r0, r1
+}
+
+// NewClient creates a new instance of Client. It also registers a testing interface on the mock and a cleanup function to assert the mocks expectations.
+// The first argument is typically a *testing.T value.
+func NewClient(t interface {
+	mock.TestingT
+	Cleanup(func())
+}) *Client {
+	mock := &Client{}
+	mock.Mock.Test(t)
+
+	t.Cleanup(func() { mock.AssertExpectations(t) })
+
+	return mock
+}

cmd/ghalistener/listener/mocks/handler.go

@@ -0,0 +1,68 @@
+// Code generated by mockery v2.36.1. DO NOT EDIT.
+
+package mocks
+
+import (
+	context "context"
+
+	actions "github.com/actions/actions-runner-controller/github/actions"
+
+	mock "github.com/stretchr/testify/mock"
+)
+
+// Handler is an autogenerated mock type for the Handler type
+type Handler struct {
+	mock.Mock
+}
+
+// HandleDesiredRunnerCount provides a mock function with given fields: ctx, count, jobsCompleted
+func (_m *Handler) HandleDesiredRunnerCount(ctx context.Context, count int, jobsCompleted int) (int, error) {
+	ret := _m.Called(ctx, count, jobsCompleted)
+
+	var r0 int
+	var r1 error
+	if rf, ok := ret.Get(0).(func(context.Context, int, int) (int, error)); ok {
+		return rf(ctx, count, jobsCompleted)
+	}
+	if rf, ok := ret.Get(0).(func(context.Context, int, int) int); ok {
+		r0 = rf(ctx, count, jobsCompleted)
+	} else {
+		r0 = ret.Get(0).(int)
+	}
+
+	if rf, ok := ret.Get(1).(func(context.Context, int, int) error); ok {
+		r1 = rf(ctx, count, jobsCompleted)
+	} else {
+		r1 = ret.Error(1)
+	}
+
+	return r0, r1
+}
+
+// HandleJobStarted provides a mock function with given fields: ctx, jobInfo
+func (_m *Handler) HandleJobStarted(ctx context.Context, jobInfo *actions.JobStarted) error {
+	ret := _m.Called(ctx, jobInfo)
+
+	var r0 error
+	if rf, ok := ret.Get(0).(func(context.Context, *actions.JobStarted) error); ok {
+		r0 = rf(ctx, jobInfo)
+	} else {
+		r0 = ret.Error(0)
+	}
+
+	return r0
+}
+
+// NewHandler creates a new instance of Handler. It also registers a testing interface on the mock and a cleanup function to assert the mocks expectations.
+// The first argument is typically a *testing.T value.
+func NewHandler(t interface {
+	mock.TestingT
+	Cleanup(func())
+}) *Handler {
+	mock := &Handler{}
+	mock.Mock.Test(t)
+
+	t.Cleanup(func() { mock.AssertExpectations(t) })
+
+	return mock
+}

cmd/ghalistener/main.go

@@ -0,0 +1,40 @@
+package main
+
+import (
+	"context"
+	"fmt"
+	"log"
+	"os"
+	"os/signal"
+	"syscall"
+
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/app"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/config"
+)
+
+func main() {
+	configPath, ok := os.LookupEnv("LISTENER_CONFIG_PATH")
+	if !ok {
+		fmt.Fprintf(os.Stderr, "Error: LISTENER_CONFIG_PATH environment variable is not set\n")
+		os.Exit(1)
+	}
+	config, err := config.Read(configPath)
+	if err != nil {
+		log.Printf("Failed to read config: %v", err)
+		os.Exit(1)
+	}
+
+	app, err := app.New(config)
+	if err != nil {
+		log.Printf("Failed to initialize app: %v", err)
+		os.Exit(1)
+	}
+
+	ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
+	defer stop()
+
+	if err := app.Run(ctx); err != nil {
+		log.Printf("Application returned an error: %v", err)
+		os.Exit(1)
+	}
+}

cmd/ghalistener/metrics/metrics.go

@@ -0,0 +1,392 @@
+package metrics
+
+import (
+	"context"
+	"net/http"
+	"strconv"
+	"time"
+
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/go-logr/logr"
+	"github.com/prometheus/client_golang/prometheus"
+	"github.com/prometheus/client_golang/prometheus/promhttp"
+)
+
+const (
+	labelKeyRunnerScaleSetName      = "name"
+	labelKeyRunnerScaleSetNamespace = "namespace"
+	labelKeyEnterprise              = "enterprise"
+	labelKeyOrganization            = "organization"
+	labelKeyRepository              = "repository"
+	labelKeyJobName                 = "job_name"
+	labelKeyJobWorkflowRef          = "job_workflow_ref"
+	labelKeyEventName               = "event_name"
+	labelKeyJobResult               = "job_result"
+	labelKeyRunnerID                = "runner_id"
+	labelKeyRunnerName              = "runner_name"
+)
+
+const githubScaleSetSubsystem = "gha"
+
+// labels
+var (
+	scaleSetLabels = []string{
+		labelKeyRunnerScaleSetName,
+		labelKeyRepository,
+		labelKeyOrganization,
+		labelKeyEnterprise,
+		labelKeyRunnerScaleSetNamespace,
+	}
+
+	jobLabels = []string{
+		labelKeyRepository,
+		labelKeyOrganization,
+		labelKeyEnterprise,
+		labelKeyJobName,
+		labelKeyJobWorkflowRef,
+		labelKeyEventName,
+	}
+
+	completedJobsTotalLabels   = append(jobLabels, labelKeyJobResult, labelKeyRunnerID, labelKeyRunnerName)
+	jobExecutionDurationLabels = append(jobLabels, labelKeyJobResult, labelKeyRunnerID, labelKeyRunnerName)
+	startedJobsTotalLabels     = append(jobLabels, labelKeyRunnerID, labelKeyRunnerName)
+	jobStartupDurationLabels   = append(jobLabels, labelKeyRunnerID, labelKeyRunnerName)
+)
+
+var (
+	assignedJobs = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "assigned_jobs",
+			Help:      "Number of jobs assigned to this scale set.",
+		},
+		scaleSetLabels,
+	)
+
+	runningJobs = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "running_jobs",
+			Help:      "Number of jobs running (or about to be run).",
+		},
+		scaleSetLabels,
+	)
+
+	registeredRunners = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "registered_runners",
+			Help:      "Number of runners registered by the scale set.",
+		},
+		scaleSetLabels,
+	)
+
+	busyRunners = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "busy_runners",
+			Help:      "Number of registered runners running a job.",
+		},
+		scaleSetLabels,
+	)
+
+	minRunners = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "min_runners",
+			Help:      "Minimum number of runners.",
+		},
+		scaleSetLabels,
+	)
+
+	maxRunners = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "max_runners",
+			Help:      "Maximum number of runners.",
+		},
+		scaleSetLabels,
+	)
+
+	desiredRunners = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "desired_runners",
+			Help:      "Number of runners desired by the scale set.",
+		},
+		scaleSetLabels,
+	)
+
+	idleRunners = prometheus.NewGaugeVec(
+		prometheus.GaugeOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "idle_runners",
+			Help:      "Number of registered runners not running a job.",
+		},
+		scaleSetLabels,
+	)
+
+	startedJobsTotal = prometheus.NewCounterVec(
+		prometheus.CounterOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "started_jobs_total",
+			Help:      "Total number of jobs started.",
+		},
+		startedJobsTotalLabels,
+	)
+
+	completedJobsTotal = prometheus.NewCounterVec(
+		prometheus.CounterOpts{
+			Name:      "completed_jobs_total",
+			Help:      "Total number of jobs completed.",
+			Subsystem: githubScaleSetSubsystem,
+		},
+		completedJobsTotalLabels,
+	)
+
+	jobStartupDurationSeconds = prometheus.NewHistogramVec(
+		prometheus.HistogramOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "job_startup_duration_seconds",
+			Help:      "Time spent waiting for workflow job to get started on the runner owned by the scale set (in seconds).",
+			Buckets:   runtimeBuckets,
+		},
+		jobStartupDurationLabels,
+	)
+
+	jobExecutionDurationSeconds = prometheus.NewHistogramVec(
+		prometheus.HistogramOpts{
+			Subsystem: githubScaleSetSubsystem,
+			Name:      "job_execution_duration_seconds",
+			Help:      "Time spent executing workflow jobs by the scale set (in seconds).",
+			Buckets:   runtimeBuckets,
+		},
+		jobExecutionDurationLabels,
+	)
+)
+
+var runtimeBuckets []float64 = []float64{
+	0.01,
+	0.05,
+	0.1,
+	0.5,
+	1,
+	2,
+	3,
+	4,
+	5,
+	6,
+	7,
+	8,
+	9,
+	10,
+	12,
+	15,
+	18,
+	20,
+	25,
+	30,
+	40,
+	50,
+	60,
+	70,
+	80,
+	90,
+	100,
+	110,
+	120,
+	150,
+	180,
+	210,
+	240,
+	300,
+	360,
+	420,
+	480,
+	540,
+	600,
+	900,
+	1200,
+	1800,
+	2400,
+	3000,
+	3600,
+}
+
+type baseLabels struct {
+	scaleSetName      string
+	scaleSetNamespace string
+	enterprise        string
+	organization      string
+	repository        string
+}
+
+func (b *baseLabels) jobLabels(jobBase *actions.JobMessageBase) prometheus.Labels {
+	return prometheus.Labels{
+		labelKeyEnterprise:     b.enterprise,
+		labelKeyOrganization:   jobBase.OwnerName,
+		labelKeyRepository:     jobBase.RepositoryName,
+		labelKeyJobName:        jobBase.JobDisplayName,
+		labelKeyJobWorkflowRef: jobBase.JobWorkflowRef,
+		labelKeyEventName:      jobBase.EventName,
+	}
+}
+
+func (b *baseLabels) scaleSetLabels() prometheus.Labels {
+	return prometheus.Labels{
+		labelKeyRunnerScaleSetName:      b.scaleSetName,
+		labelKeyRunnerScaleSetNamespace: b.scaleSetNamespace,
+		labelKeyEnterprise:              b.enterprise,
+		labelKeyOrganization:            b.organization,
+		labelKeyRepository:              b.repository,
+	}
+}
+
+func (b *baseLabels) completedJobLabels(msg *actions.JobCompleted) prometheus.Labels {
+	l := b.jobLabels(&msg.JobMessageBase)
+	l[labelKeyRunnerID] = strconv.Itoa(msg.RunnerId)
+	l[labelKeyJobResult] = msg.Result
+	l[labelKeyRunnerName] = msg.RunnerName
+	return l
+}
+
+func (b *baseLabels) startedJobLabels(msg *actions.JobStarted) prometheus.Labels {
+	l := b.jobLabels(&msg.JobMessageBase)
+	l[labelKeyRunnerID] = strconv.Itoa(msg.RunnerId)
+	l[labelKeyRunnerName] = msg.RunnerName
+	return l
+}
+
+//go:generate mockery --name Publisher --output ./mocks --outpkg mocks --case underscore
+type Publisher interface {
+	PublishStatic(min, max int)
+	PublishStatistics(stats *actions.RunnerScaleSetStatistic)
+	PublishJobStarted(msg *actions.JobStarted)
+	PublishJobCompleted(msg *actions.JobCompleted)
+	PublishDesiredRunners(count int)
+}
+
+//go:generate mockery --name ServerPublisher --output ./mocks --outpkg mocks --case underscore
+type ServerPublisher interface {
+	Publisher
+	ListenAndServe(ctx context.Context) error
+}
+
+var (
+	_ Publisher       = &discard{}
+	_ ServerPublisher = &exporter{}
+)
+
+var Discard Publisher = &discard{}
+
+type exporter struct {
+	logger logr.Logger
+	baseLabels
+	srv *http.Server
+}
+
+type ExporterConfig struct {
+	ScaleSetName      string
+	ScaleSetNamespace string
+	Enterprise        string
+	Organization      string
+	Repository        string
+	ServerAddr        string
+	ServerEndpoint    string
+	Logger            logr.Logger
+}
+
+func NewExporter(config ExporterConfig) ServerPublisher {
+	reg := prometheus.NewRegistry()
+	reg.MustRegister(
+		assignedJobs,
+		runningJobs,
+		registeredRunners,
+		busyRunners,
+		minRunners,
+		maxRunners,
+		desiredRunners,
+		idleRunners,
+		startedJobsTotal,
+		completedJobsTotal,
+		jobStartupDurationSeconds,
+		jobExecutionDurationSeconds,
+	)
+
+	mux := http.NewServeMux()
+	mux.Handle(
+		config.ServerEndpoint,
+		promhttp.HandlerFor(reg, promhttp.HandlerOpts{Registry: reg}),
+	)
+
+	return &exporter{
+		logger: config.Logger.WithName("metrics"),
+		baseLabels: baseLabels{
+			scaleSetName:      config.ScaleSetName,
+			scaleSetNamespace: config.ScaleSetNamespace,
+			enterprise:        config.Enterprise,
+			organization:      config.Organization,
+			repository:        config.Repository,
+		},
+		srv: &http.Server{
+			Addr:    config.ServerAddr,
+			Handler: mux,
+		},
+	}
+}
+
+func (e *exporter) ListenAndServe(ctx context.Context) error {
+	e.logger.Info("starting metrics server", "addr", e.srv.Addr)
+	go func() {
+		<-ctx.Done()
+		e.logger.Info("stopping metrics server", "err", ctx.Err())
+		ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
+		defer cancel()
+		e.srv.Shutdown(ctx)
+	}()
+	return e.srv.ListenAndServe()
+}
+
+func (m *exporter) PublishStatic(min, max int) {
+	l := m.scaleSetLabels()
+	maxRunners.With(l).Set(float64(max))
+	minRunners.With(l).Set(float64(min))
+}
+
+func (e *exporter) PublishStatistics(stats *actions.RunnerScaleSetStatistic) {
+	l := e.scaleSetLabels()
+
+	assignedJobs.With(l).Set(float64(stats.TotalAssignedJobs))
+	runningJobs.With(l).Set(float64(stats.TotalRunningJobs))
+	registeredRunners.With(l).Set(float64(stats.TotalRegisteredRunners))
+	busyRunners.With(l).Set(float64(stats.TotalBusyRunners))
+	idleRunners.With(l).Set(float64(stats.TotalIdleRunners))
+}
+
+func (e *exporter) PublishJobStarted(msg *actions.JobStarted) {
+	l := e.startedJobLabels(msg)
+	startedJobsTotal.With(l).Inc()
+
+	startupDuration := msg.JobMessageBase.RunnerAssignTime.Unix() - msg.JobMessageBase.ScaleSetAssignTime.Unix()
+	jobStartupDurationSeconds.With(l).Observe(float64(startupDuration))
+}
+
+func (e *exporter) PublishJobCompleted(msg *actions.JobCompleted) {
+	l := e.completedJobLabels(msg)
+	completedJobsTotal.With(l).Inc()
+
+	executionDuration := msg.JobMessageBase.FinishTime.Unix() - msg.JobMessageBase.RunnerAssignTime.Unix()
+	jobExecutionDurationSeconds.With(l).Observe(float64(executionDuration))
+}
+
+func (m *exporter) PublishDesiredRunners(count int) {
+	desiredRunners.With(m.scaleSetLabels()).Set(float64(count))
+}
+
+type discard struct{}
+
+func (*discard) PublishStatic(int, int)                             {}
+func (*discard) PublishStatistics(*actions.RunnerScaleSetStatistic) {}
+func (*discard) PublishJobStarted(*actions.JobStarted)              {}
+func (*discard) PublishJobCompleted(*actions.JobCompleted)          {}
+func (*discard) PublishDesiredRunners(int)                          {}

cmd/ghalistener/metrics/mocks/publisher.go

@@ -0,0 +1,53 @@
+// Code generated by mockery v2.36.1. DO NOT EDIT.
+
+package mocks
+
+import (
+	actions "github.com/actions/actions-runner-controller/github/actions"
+
+	mock "github.com/stretchr/testify/mock"
+)
+
+// Publisher is an autogenerated mock type for the Publisher type
+type Publisher struct {
+	mock.Mock
+}
+
+// PublishDesiredRunners provides a mock function with given fields: count
+func (_m *Publisher) PublishDesiredRunners(count int) {
+	_m.Called(count)
+}
+
+// PublishJobCompleted provides a mock function with given fields: msg
+func (_m *Publisher) PublishJobCompleted(msg *actions.JobCompleted) {
+	_m.Called(msg)
+}
+
+// PublishJobStarted provides a mock function with given fields: msg
+func (_m *Publisher) PublishJobStarted(msg *actions.JobStarted) {
+	_m.Called(msg)
+}
+
+// PublishStatic provides a mock function with given fields: min, max
+func (_m *Publisher) PublishStatic(min int, max int) {
+	_m.Called(min, max)
+}
+
+// PublishStatistics provides a mock function with given fields: stats
+func (_m *Publisher) PublishStatistics(stats *actions.RunnerScaleSetStatistic) {
+	_m.Called(stats)
+}
+
+// NewPublisher creates a new instance of Publisher. It also registers a testing interface on the mock and a cleanup function to assert the mocks expectations.
+// The first argument is typically a *testing.T value.
+func NewPublisher(t interface {
+	mock.TestingT
+	Cleanup(func())
+}) *Publisher {
+	mock := &Publisher{}
+	mock.Mock.Test(t)
+
+	t.Cleanup(func() { mock.AssertExpectations(t) })
+
+	return mock
+}

cmd/ghalistener/metrics/mocks/server_publisher.go

@@ -0,0 +1,69 @@
+// Code generated by mockery v2.36.1. DO NOT EDIT.
+
+package mocks
+
+import (
+	context "context"
+
+	actions "github.com/actions/actions-runner-controller/github/actions"
+
+	mock "github.com/stretchr/testify/mock"
+)
+
+// ServerPublisher is an autogenerated mock type for the ServerPublisher type
+type ServerPublisher struct {
+	mock.Mock
+}
+
+// ListenAndServe provides a mock function with given fields: ctx
+func (_m *ServerPublisher) ListenAndServe(ctx context.Context) error {
+	ret := _m.Called(ctx)
+
+	var r0 error
+	if rf, ok := ret.Get(0).(func(context.Context) error); ok {
+		r0 = rf(ctx)
+	} else {
+		r0 = ret.Error(0)
+	}
+
+	return r0
+}
+
+// PublishDesiredRunners provides a mock function with given fields: count
+func (_m *ServerPublisher) PublishDesiredRunners(count int) {
+	_m.Called(count)
+}
+
+// PublishJobCompleted provides a mock function with given fields: msg
+func (_m *ServerPublisher) PublishJobCompleted(msg *actions.JobCompleted) {
+	_m.Called(msg)
+}
+
+// PublishJobStarted provides a mock function with given fields: msg
+func (_m *ServerPublisher) PublishJobStarted(msg *actions.JobStarted) {
+	_m.Called(msg)
+}
+
+// PublishStatic provides a mock function with given fields: min, max
+func (_m *ServerPublisher) PublishStatic(min int, max int) {
+	_m.Called(min, max)
+}
+
+// PublishStatistics provides a mock function with given fields: stats
+func (_m *ServerPublisher) PublishStatistics(stats *actions.RunnerScaleSetStatistic) {
+	_m.Called(stats)
+}
+
+// NewServerPublisher creates a new instance of ServerPublisher. It also registers a testing interface on the mock and a cleanup function to assert the mocks expectations.
+// The first argument is typically a *testing.T value.
+func NewServerPublisher(t interface {
+	mock.TestingT
+	Cleanup(func())
+}) *ServerPublisher {
+	mock := &ServerPublisher{}
+	mock.Mock.Test(t)
+
+	t.Cleanup(func() { mock.AssertExpectations(t) })
+
+	return mock
+}

cmd/ghalistener/worker/worker.go

@@ -0,0 +1,242 @@
+package worker
+
+import (
+	"context"
+	"encoding/json"
+	"fmt"
+
+	"github.com/actions/actions-runner-controller/apis/actions.github.com/v1alpha1"
+	"github.com/actions/actions-runner-controller/cmd/ghalistener/listener"
+	"github.com/actions/actions-runner-controller/github/actions"
+	"github.com/actions/actions-runner-controller/logging"
+	jsonpatch "github.com/evanphx/json-patch"
+	"github.com/go-logr/logr"
+	kerrors "k8s.io/apimachinery/pkg/api/errors"
+	"k8s.io/apimachinery/pkg/types"
+	"k8s.io/client-go/kubernetes"
+	"k8s.io/client-go/rest"
+)
+
+const workerName = "kubernetesworker"
+
+type Option func(*Worker)
+
+func WithLogger(logger logr.Logger) Option {
+	return func(w *Worker) {
+		logger = logger.WithName(workerName)
+		w.logger = &logger
+	}
+}
+
+type Config struct {
+	EphemeralRunnerSetNamespace string
+	EphemeralRunnerSetName      string
+	MaxRunners                  int
+	MinRunners                  int
+}
+
+// The Worker's role is to process the messages it receives from the listener.
+// It then initiates Kubernetes API requests to carry out the necessary actions.
+type Worker struct {
+	clientset   *kubernetes.Clientset
+	config      Config
+	lastPatch   int
+	lastPatchID int
+	logger      *logr.Logger
+}
+
+var _ listener.Handler = (*Worker)(nil)
+
+func New(config Config, options ...Option) (*Worker, error) {
+	w := &Worker{
+		config:      config,
+		lastPatch:   -1,
+		lastPatchID: -1,
+	}
+
+	conf, err := rest.InClusterConfig()
+	if err != nil {
+		return nil, err
+	}
+
+	clientset, err := kubernetes.NewForConfig(conf)
+	if err != nil {
+		return nil, err
+	}
+
+	w.clientset = clientset
+
+	for _, option := range options {
+		option(w)
+	}
+
+	if err := w.applyDefaults(); err != nil {
+		return nil, err
+	}
+
+	return w, nil
+}
+
+func (w *Worker) applyDefaults() error {
+	if w.logger == nil {
+		logger, err := logging.NewLogger(logging.LogLevelDebug, logging.LogFormatJSON)
+		if err != nil {
+			return fmt.Errorf("NewLogger failed: %w", err)
+		}
+		logger = logger.WithName(workerName)
+		w.logger = &logger
+	}
+
+	return nil
+}
+
+// HandleJobStarted updates the job information for the ephemeral runner when a job is started.
+// It takes a context and a jobInfo parameter which contains the details of the started job.
+// This update marks the ephemeral runner so that the controller would have more context
+// about the ephemeral runner that should not be deleted when scaling down.
+// It returns an error if there is any issue with updating the job information.
+func (w *Worker) HandleJobStarted(ctx context.Context, jobInfo *actions.JobStarted) error {
+	w.logger.Info("Updating job info for the runner",
+		"runnerName", jobInfo.RunnerName,
+		"ownerName", jobInfo.OwnerName,
+		"repoName", jobInfo.RepositoryName,
+		"workflowRef", jobInfo.JobWorkflowRef,
+		"workflowRunId", jobInfo.WorkflowRunId,
+		"jobDisplayName", jobInfo.JobDisplayName,
+		"requestId", jobInfo.RunnerRequestId)
+
+	original, err := json.Marshal(&v1alpha1.EphemeralRunner{})
+	if err != nil {
+		return fmt.Errorf("failed to marshal empty ephemeral runner: %w", err)
+	}
+
+	patch, err := json.Marshal(
+		&v1alpha1.EphemeralRunner{
+			Status: v1alpha1.EphemeralRunnerStatus{
+				JobRequestId:      jobInfo.RunnerRequestId,
+				JobRepositoryName: fmt.Sprintf("%s/%s", jobInfo.OwnerName, jobInfo.RepositoryName),
+				WorkflowRunId:     jobInfo.WorkflowRunId,
+				JobWorkflowRef:    jobInfo.JobWorkflowRef,
+				JobDisplayName:    jobInfo.JobDisplayName,
+			},
+		},
+	)
+	if err != nil {
+		return fmt.Errorf("failed to marshal ephemeral runner patch: %w", err)
+	}
+
+	mergePatch, err := jsonpatch.CreateMergePatch(original, patch)
+	if err != nil {
+		return fmt.Errorf("failed to create merge patch json for ephemeral runner: %w", err)
+	}
+
+	w.logger.Info("Updating ephemeral runner with merge patch", "json", string(mergePatch))
+
+	patchedStatus := &v1alpha1.EphemeralRunner{}
+	err = w.clientset.RESTClient().
+		Patch(types.MergePatchType).
+		Prefix("apis", v1alpha1.GroupVersion.Group, v1alpha1.GroupVersion.Version).
+		Namespace(w.config.EphemeralRunnerSetNamespace).
+		Resource("EphemeralRunners").
+		Name(jobInfo.RunnerName).
+		SubResource("status").
+		Body(mergePatch).
+		Do(ctx).
+		Into(patchedStatus)
+	if err != nil {
+		if kerrors.IsNotFound(err) {
+			w.logger.Info("Ephemeral runner not found, skipping patching of ephemeral runner status", "runnerName", jobInfo.RunnerName)
+			return nil
+		}
+		return fmt.Errorf("could not patch ephemeral runner status, patch JSON: %s, error: %w", string(mergePatch), err)
+	}
+
+	w.logger.Info("Ephemeral runner status updated with the merge patch successfully.")
+
+	return nil
+}
+
+// HandleDesiredRunnerCount handles the desired runner count by scaling the ephemeral runner set.
+// The function calculates the target runner count based on the minimum and maximum runner count configuration.
+// If the target runner count is the same as the last patched count, it skips patching and returns nil.
+// Otherwise, it creates a merge patch JSON for updating the ephemeral runner set with the desired count.
+// The function then scales the ephemeral runner set by applying the merge patch.
+// Finally, it logs the scaled ephemeral runner set details and returns nil if successful.
+// If any error occurs during the process, it returns an error with a descriptive message.
+func (w *Worker) HandleDesiredRunnerCount(ctx context.Context, count int, jobsCompleted int) (int, error) {
+	// Max runners should always be set by the resource builder either to the configured value,
+	// or the maximum int32 (resourcebuilder.newAutoScalingListener()).
+	targetRunnerCount := min(w.config.MinRunners+count, w.config.MaxRunners)
+
+	logValues := []any{
+		"assigned job", count,
+		"decision", targetRunnerCount,
+		"min", w.config.MinRunners,
+		"max", w.config.MaxRunners,
+		"currentRunnerCount", w.lastPatch,
+		"jobsCompleted", jobsCompleted,
+	}
+
+	if count == 0 && jobsCompleted == 0 {
+		w.lastPatchID = 0
+	} else {
+		w.lastPatchID++
+	}
+
+	w.lastPatch = targetRunnerCount
+
+	original, err := json.Marshal(
+		&v1alpha1.EphemeralRunnerSet{
+			Spec: v1alpha1.EphemeralRunnerSetSpec{
+				Replicas: -1,
+				PatchID:  -1,
+			},
+		},
+	)
+	if err != nil {
+		return 0, fmt.Errorf("failed to marshal empty ephemeral runner set: %w", err)
+	}
+
+	patch, err := json.Marshal(
+		&v1alpha1.EphemeralRunnerSet{
+			Spec: v1alpha1.EphemeralRunnerSetSpec{
+				Replicas: targetRunnerCount,
+				PatchID:  w.lastPatchID,
+			},
+		},
+	)
+	if err != nil {
+		w.logger.Error(err, "could not marshal patch ephemeral runner set")
+		return 0, err
+	}
+
+	mergePatch, err := jsonpatch.CreateMergePatch(original, patch)
+	if err != nil {
+		return 0, fmt.Errorf("failed to create merge patch json for ephemeral runner set: %w", err)
+	}
+
+	w.logger.Info("Created merge patch json for EphemeralRunnerSet update", "json", string(mergePatch))
+
+	w.logger.Info("Scaling ephemeral runner set", logValues...)
+
+	patchedEphemeralRunnerSet := &v1alpha1.EphemeralRunnerSet{}
+	err = w.clientset.RESTClient().
+		Patch(types.MergePatchType).
+		Prefix("apis", v1alpha1.GroupVersion.Group, v1alpha1.GroupVersion.Version).
+		Namespace(w.config.EphemeralRunnerSetNamespace).
+		Resource("ephemeralrunnersets").
+		Name(w.config.EphemeralRunnerSetName).
+		Body([]byte(mergePatch)).
+		Do(ctx).
+		Into(patchedEphemeralRunnerSet)
+	if err != nil {
+		return 0, fmt.Errorf("could not patch ephemeral runner set , patch JSON: %s, error: %w", string(mergePatch), err)
+	}
+
+	w.logger.Info("Ephemeral runner set scaled.",
+		"namespace", w.config.EphemeralRunnerSetNamespace,
+		"name", w.config.EphemeralRunnerSetName,
+		"replicas", patchedEphemeralRunnerSet.Spec.Replicas,
+	)
+	return targetRunnerCount, nil
+}

cmd/githubrunnerscalesetlistener/autoScalerMessageListener.go

@@ -129,7 +129,7 @@ func (m *AutoScalerClient) Close() error {
 	return m.client.Close()
 }
 
-func (m *AutoScalerClient) GetRunnerScaleSetMessage(ctx context.Context, handler func(msg *actions.RunnerScaleSetMessage) error) error {
+func (m *AutoScalerClient) GetRunnerScaleSetMessage(ctx context.Context, handler func(msg *actions.RunnerScaleSetMessage) error, maxCapacity int) error {
 	if m.initialMessage != nil {
 		err := handler(m.initialMessage)
 		if err != nil {
@@ -141,7 +141,7 @@ func (m *AutoScalerClient) GetRunnerScaleSetMessage(ctx context.Context, handler
 	}
 
 	for {
-		message, err := m.client.GetMessage(ctx, m.lastMessageId)
+		message, err := m.client.GetMessage(ctx, m.lastMessageId, maxCapacity)
 		if err != nil {
 			return fmt.Errorf("get message failed from refreshing client. %w", err)
 		}

cmd/githubrunnerscalesetlistener/autoScalerMessageListener_test.go

@@ -317,7 +317,7 @@ func TestGetRunnerScaleSetMessage(t *testing.T) {
 		Statistics: &actions.RunnerScaleSetStatistic{},
 	}
 	mockActionsClient.On("CreateMessageSession", ctx, 1, mock.Anything).Return(session, nil)
-	mockSessionClient.On("GetMessage", ctx, int64(0)).Return(&actions.RunnerScaleSetMessage{
+	mockSessionClient.On("GetMessage", ctx, int64(0), mock.Anything).Return(&actions.RunnerScaleSetMessage{
 		MessageId:   1,
 		MessageType: "test",
 		Body:        "test",
@@ -332,7 +332,7 @@ func TestGetRunnerScaleSetMessage(t *testing.T) {
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 
 	assert.NoError(t, err, "Error getting message")
 	assert.Equal(t, int64(0), asClient.lastMessageId, "Initial message")
@@ -340,7 +340,7 @@ func TestGetRunnerScaleSetMessage(t *testing.T) {
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 
 	assert.NoError(t, err, "Error getting message")
 	assert.Equal(t, int64(1), asClient.lastMessageId, "Last message id should be updated")
@@ -368,7 +368,7 @@ func TestGetRunnerScaleSetMessage_HandleFailed(t *testing.T) {
 		Statistics: &actions.RunnerScaleSetStatistic{},
 	}
 	mockActionsClient.On("CreateMessageSession", ctx, 1, mock.Anything).Return(session, nil)
-	mockSessionClient.On("GetMessage", ctx, int64(0)).Return(&actions.RunnerScaleSetMessage{
+	mockSessionClient.On("GetMessage", ctx, int64(0), mock.Anything).Return(&actions.RunnerScaleSetMessage{
 		MessageId:   1,
 		MessageType: "test",
 		Body:        "test",
@@ -383,14 +383,14 @@ func TestGetRunnerScaleSetMessage_HandleFailed(t *testing.T) {
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 
 	assert.NoError(t, err, "Error getting message")
 
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return fmt.Errorf("error")
-	})
+	}, 10)
 
 	assert.ErrorContains(t, err, "handle message failed. error", "Error getting message")
 	assert.Equal(t, int64(0), asClient.lastMessageId, "Last message id should not be updated")
@@ -419,7 +419,7 @@ func TestGetRunnerScaleSetMessage_HandleInitialMessage(t *testing.T) {
 			TotalAssignedJobs:  2,
 		},
 	}
-	mockActionsClient.On("CreateMessageSession", ctx, 1, mock.Anything).Return(session, nil)
+	mockActionsClient.On("CreateMessageSession", ctx, 1, mock.Anything, mock.Anything).Return(session, nil)
 	mockActionsClient.On("GetAcquirableJobs", ctx, 1).Return(&actions.AcquirableJobList{
 		Count: 1,
 		Jobs: []actions.AcquirableJob{
@@ -439,7 +439,7 @@ func TestGetRunnerScaleSetMessage_HandleInitialMessage(t *testing.T) {
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 
 	assert.NoError(t, err, "Error getting message")
 	assert.Nil(t, asClient.initialMessage, "Initial message should be nil")
@@ -488,7 +488,7 @@ func TestGetRunnerScaleSetMessage_HandleInitialMessageFailed(t *testing.T) {
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return fmt.Errorf("error")
-	})
+	}, 10)
 
 	assert.ErrorContains(t, err, "fail to process initial message. error", "Error getting message")
 	assert.NotNil(t, asClient.initialMessage, "Initial message should be nil")
@@ -516,8 +516,8 @@ func TestGetRunnerScaleSetMessage_RetryUntilGetMessage(t *testing.T) {
 		Statistics: &actions.RunnerScaleSetStatistic{},
 	}
 	mockActionsClient.On("CreateMessageSession", ctx, 1, mock.Anything).Return(session, nil)
-	mockSessionClient.On("GetMessage", ctx, int64(0)).Return(nil, nil).Times(3)
-	mockSessionClient.On("GetMessage", ctx, int64(0)).Return(&actions.RunnerScaleSetMessage{
+	mockSessionClient.On("GetMessage", ctx, int64(0), mock.Anything).Return(nil, nil).Times(3)
+	mockSessionClient.On("GetMessage", ctx, int64(0), mock.Anything).Return(&actions.RunnerScaleSetMessage{
 		MessageId:   1,
 		MessageType: "test",
 		Body:        "test",
@@ -532,13 +532,13 @@ func TestGetRunnerScaleSetMessage_RetryUntilGetMessage(t *testing.T) {
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 	assert.NoError(t, err, "Error getting initial message")
 
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 
 	assert.NoError(t, err, "Error getting message")
 	assert.Equal(t, int64(1), asClient.lastMessageId, "Last message id should be updated")
@@ -565,7 +565,7 @@ func TestGetRunnerScaleSetMessage_ErrorOnGetMessage(t *testing.T) {
 		Statistics: &actions.RunnerScaleSetStatistic{},
 	}
 	mockActionsClient.On("CreateMessageSession", ctx, 1, mock.Anything).Return(session, nil)
-	mockSessionClient.On("GetMessage", ctx, int64(0)).Return(nil, fmt.Errorf("error"))
+	mockSessionClient.On("GetMessage", ctx, int64(0), mock.Anything).Return(nil, fmt.Errorf("error"))
 
 	asClient, err := NewAutoScalerClient(ctx, mockActionsClient, &logger, 1, func(asc *AutoScalerClient) {
 		asc.client = mockSessionClient
@@ -575,12 +575,12 @@ func TestGetRunnerScaleSetMessage_ErrorOnGetMessage(t *testing.T) {
 	// process initial message
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		return nil
-	})
+	}, 10)
 	assert.NoError(t, err, "Error getting initial message")
 
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		return fmt.Errorf("Should not be called")
-	})
+	}, 10)
 
 	assert.ErrorContains(t, err, "get message failed from refreshing client. error", "Error should be returned")
 	assert.Equal(t, int64(0), asClient.lastMessageId, "Last message id should be updated")
@@ -608,7 +608,7 @@ func TestDeleteRunnerScaleSetMessage_Error(t *testing.T) {
 		Statistics: &actions.RunnerScaleSetStatistic{},
 	}
 	mockActionsClient.On("CreateMessageSession", ctx, 1, mock.Anything).Return(session, nil)
-	mockSessionClient.On("GetMessage", ctx, int64(0)).Return(&actions.RunnerScaleSetMessage{
+	mockSessionClient.On("GetMessage", ctx, int64(0), mock.Anything).Return(&actions.RunnerScaleSetMessage{
 		MessageId:   1,
 		MessageType: "test",
 		Body:        "test",
@@ -623,13 +623,13 @@ func TestDeleteRunnerScaleSetMessage_Error(t *testing.T) {
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 	assert.NoError(t, err, "Error getting initial message")
 
 	err = asClient.GetRunnerScaleSetMessage(ctx, func(msg *actions.RunnerScaleSetMessage) error {
 		logger.Info("Message received", "messageId", msg.MessageId, "messageType", msg.MessageType, "body", msg.Body)
 		return nil
-	})
+	}, 10)
 
 	assert.ErrorContains(t, err, "delete message failed from refreshing client. error", "Error getting message")
 	assert.Equal(t, int64(1), asClient.lastMessageId, "Last message id should be updated")

cmd/githubrunnerscalesetlistener/autoScalerService.go

@@ -5,9 +5,9 @@ import (
 	"encoding/json"
 	"errors"
 	"fmt"
-	"math"
 	"strings"
 
+	"github.com/actions/actions-runner-controller/cmd/githubrunnerscalesetlistener/config"
 	"github.com/actions/actions-runner-controller/github/actions"
 	"github.com/go-logr/logr"
 )
@@ -30,7 +30,7 @@ type Service struct {
 	errs               []error
 }
 
-func WithPrometheusMetrics(conf RunnerScaleSetListenerConfig) func(*Service) {
+func WithPrometheusMetrics(conf config.Config) func(*Service) {
 	return func(svc *Service) {
 		parsedURL, err := actions.ParseGitHubConfigFromURL(conf.ConfigureUrl)
 		if err != nil {
@@ -81,6 +81,7 @@ func NewService(
 }
 
 func (s *Service) Start() error {
+	s.metricsExporter.publishStatic(s.settings.MaxRunners, s.settings.MinRunners)
 	for {
 		s.logger.Info("waiting for message...")
 		select {
@@ -88,7 +89,7 @@ func (s *Service) Start() error {
 			s.logger.Info("service is stopped.")
 			return nil
 		default:
-			err := s.rsClient.GetRunnerScaleSetMessage(s.ctx, s.processMessage)
+			err := s.rsClient.GetRunnerScaleSetMessage(s.ctx, s.processMessage, s.settings.MaxRunners)
 			if err != nil {
 				return fmt.Errorf("could not get and process message. %w", err)
 			}
@@ -204,7 +205,9 @@ func (s *Service) processMessage(message *actions.RunnerScaleSetMessage) error {
 }
 
 func (s *Service) scaleForAssignedJobCount(count int) error {
-	targetRunnerCount := int(math.Max(math.Min(float64(s.settings.MaxRunners), float64(count)), float64(s.settings.MinRunners)))
+	// Max runners should always be set by the resource builder either to the configured value,
+	// or the maximum int32 (resourcebuilder.newAutoScalingListener()).
+	targetRunnerCount := min(s.settings.MinRunners+count, s.settings.MaxRunners)
 	s.metricsExporter.publishDesiredRunners(targetRunnerCount)
 	if targetRunnerCount != s.currentRunnerCount {
 		s.logger.Info("try scale runner request up/down base on assigned job count",

cmd/githubrunnerscalesetlistener/autoScalerService_test.go

@@ -64,7 +64,7 @@ func TestStart(t *testing.T) {
 	)
 	require.NoError(t, err)
 
-	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything).Run(func(args mock.Arguments) { cancel() }).Return(nil).Once()
+	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything, mock.Anything).Run(func(mock.Arguments) { cancel() }).Return(nil).Once()
 
 	err = service.Start()
 
@@ -98,7 +98,7 @@ func TestStart_ScaleToMinRunners(t *testing.T) {
 	)
 	require.NoError(t, err)
 
-	mockRsClient.On("GetRunnerScaleSetMessage", ctx, mock.Anything).Run(func(args mock.Arguments) {
+	mockRsClient.On("GetRunnerScaleSetMessage", ctx, mock.Anything, mock.Anything).Run(func(args mock.Arguments) {
 		_ = service.scaleForAssignedJobCount(5)
 	}).Return(nil)
 
@@ -137,7 +137,7 @@ func TestStart_ScaleToMinRunnersFailed(t *testing.T) {
 	require.NoError(t, err)
 
 	c := mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 5).Return(fmt.Errorf("error")).Once()
-	mockRsClient.On("GetRunnerScaleSetMessage", ctx, mock.Anything).Run(func(args mock.Arguments) {
+	mockRsClient.On("GetRunnerScaleSetMessage", ctx, mock.Anything, mock.Anything).Run(func(args mock.Arguments) {
 		_ = service.scaleForAssignedJobCount(5)
 	}).Return(c.ReturnArguments.Get(0))
 
@@ -172,8 +172,8 @@ func TestStart_GetMultipleMessages(t *testing.T) {
 	)
 	require.NoError(t, err)
 
-	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything).Return(nil).Times(5)
-	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything).Run(func(args mock.Arguments) { cancel() }).Return(nil).Once()
+	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything, mock.Anything).Return(nil).Times(5)
+	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything, mock.Anything).Run(func(args mock.Arguments) { cancel() }).Return(nil).Once()
 
 	err = service.Start()
 
@@ -207,8 +207,8 @@ func TestStart_ErrorOnMessage(t *testing.T) {
 	)
 	require.NoError(t, err)
 
-	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything).Return(nil).Times(2)
-	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything).Return(fmt.Errorf("error")).Once()
+	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything, mock.Anything).Return(nil).Times(2)
+	mockRsClient.On("GetRunnerScaleSetMessage", service.ctx, mock.Anything, mock.Anything).Return(fmt.Errorf("error")).Once()
 
 	err = service.Start()
 
@@ -397,7 +397,7 @@ func TestProcessMessage_MultipleMessages(t *testing.T) {
 	require.NoError(t, err)
 
 	mockRsClient.On("AcquireJobsForRunnerScaleSet", ctx, mock.MatchedBy(func(ids []int64) bool { return ids[0] == 3 && ids[1] == 4 })).Return(nil).Once()
-	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 2).Run(func(args mock.Arguments) { cancel() }).Return(nil).Once()
+	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 3).Run(func(args mock.Arguments) { cancel() }).Return(nil).Once()
 
 	err = service.processMessage(&actions.RunnerScaleSetMessage{
 		MessageId:   1,
@@ -523,9 +523,9 @@ func TestScaleForAssignedJobCount_ScaleWithinMinMax(t *testing.T) {
 	require.NoError(t, err)
 
 	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 1).Return(nil).Once()
-	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 3).Return(nil).Once()
+	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 4).Return(nil).Once()
 	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 5).Return(nil).Once()
-	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 1).Return(nil).Once()
+	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 2).Return(nil).Once()
 	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 5).Return(nil).Once()
 
 	err = service.scaleForAssignedJobCount(0)
@@ -569,7 +569,7 @@ func TestScaleForAssignedJobCount_ScaleFailed(t *testing.T) {
 	)
 	require.NoError(t, err)
 
-	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 2).Return(fmt.Errorf("error"))
+	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 3).Return(fmt.Errorf("error"))
 
 	err = service.scaleForAssignedJobCount(2)
 
@@ -605,8 +605,23 @@ func TestProcessMessage_JobStartedMessage(t *testing.T) {
 
 	service.currentRunnerCount = 1
 
-	mockKubeManager.On("UpdateEphemeralRunnerWithJobInfo", ctx, service.settings.Namespace, "runner1", "owner1", "repo1", ".github/workflows/ci.yaml", "job1", int64(100), int64(3)).Run(func(args mock.Arguments) { cancel() }).Return(nil).Once()
+	mockKubeManager.On(
+		"UpdateEphemeralRunnerWithJobInfo",
+		ctx,
+		service.settings.Namespace,
+		"runner1",
+		"owner1",
+		"repo1",
+		".github/workflows/ci.yaml",
+		"job1",
+		int64(100),
+		int64(3),
+	).Run(
+		func(_ mock.Arguments) { cancel() },
+	).Return(nil).Once()
+
 	mockRsClient.On("AcquireJobsForRunnerScaleSet", ctx, mock.MatchedBy(func(ids []int64) bool { return len(ids) == 0 })).Return(nil).Once()
+	mockKubeManager.On("ScaleEphemeralRunnerSet", ctx, service.settings.Namespace, service.settings.ResourceName, 2).Return(nil)
 
 	err = service.processMessage(&actions.RunnerScaleSetMessage{
 		MessageId:   1,

cmd/githubrunnerscalesetlistener/config/config.go

@@ -0,0 +1,76 @@
+package config
+
+import (
+	"encoding/json"
+	"fmt"
+	"os"
+)
+
+type Config struct {
+	ConfigureUrl                string `json:"configureUrl"`
+	AppID                       int64  `json:"appID"`
+	AppInstallationID           int64  `json:"appInstallationID"`
+	AppPrivateKey               string `json:"appPrivateKey"`
+	Token                       string `json:"token"`
+	EphemeralRunnerSetNamespace string `json:"ephemeralRunnerSetNamespace"`
+	EphemeralRunnerSetName      string `json:"ephemeralRunnerSetName"`
+	MaxRunners                  int    `json:"maxRunners"`
+	MinRunners                  int    `json:"minRunners"`
+	RunnerScaleSetId            int    `json:"runnerScaleSetId"`
+	RunnerScaleSetName          string `json:"runnerScaleSetName"`
+	ServerRootCA                string `json:"serverRootCA"`
+	LogLevel                    string `json:"logLevel"`
+	LogFormat                   string `json:"logFormat"`
+	MetricsAddr                 string `json:"metricsAddr"`
+	MetricsEndpoint             string `json:"metricsEndpoint"`
+}
+
+func Read(path string) (Config, error) {
+	f, err := os.Open(path)
+	if err != nil {
+		return Config{}, err
+	}
+	defer f.Close()
+
+	var config Config
+	if err := json.NewDecoder(f).Decode(&config); err != nil {
+		return Config{}, fmt.Errorf("failed to decode config: %w", err)
+	}
+
+	if err := config.validate(); err != nil {
+		return Config{}, fmt.Errorf("failed to validate config: %w", err)
+	}
+
+	return config, nil
+}
+
+func (c *Config) validate() error {
+	if len(c.ConfigureUrl) == 0 {
+		return fmt.Errorf("GitHubConfigUrl is not provided")
+	}
+
+	if len(c.EphemeralRunnerSetNamespace) == 0 || len(c.EphemeralRunnerSetName) == 0 {
+		return fmt.Errorf("EphemeralRunnerSetNamespace '%s' or EphemeralRunnerSetName '%s' is missing", c.EphemeralRunnerSetNamespace, c.EphemeralRunnerSetName)
+	}
+
+	if c.RunnerScaleSetId == 0 {
+		return fmt.Errorf("RunnerScaleSetId '%d' is missing", c.RunnerScaleSetId)
+	}
+
+	if c.MaxRunners < c.MinRunners {
+		return fmt.Errorf("MinRunners '%d' cannot be greater than MaxRunners '%d'", c.MinRunners, c.MaxRunners)
+	}
+
+	hasToken := len(c.Token) > 0
+	hasPrivateKeyConfig := c.AppID > 0 && c.AppPrivateKey != ""
+
+	if !hasToken && !hasPrivateKeyConfig {
+		return fmt.Errorf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(c.Token), c.AppID, c.AppInstallationID, len(c.AppPrivateKey))
+	}
+
+	if hasToken && hasPrivateKeyConfig {
+		return fmt.Errorf("only one GitHub auth method supported at a time. Have both PAT and App auth: token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(c.Token), c.AppID, c.AppInstallationID, len(c.AppPrivateKey))
+	}
+
+	return nil
+}

cmd/githubrunnerscalesetlistener/config/config_test.go

@@ -0,0 +1,92 @@
+package config
+
+import (
+	"fmt"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+)
+
+func TestConfigValidationMinMax(t *testing.T) {
+	config := &Config{
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+		MinRunners:                  5,
+		MaxRunners:                  2,
+		Token:                       "token",
+	}
+	err := config.validate()
+	assert.ErrorContains(t, err, "MinRunners '5' cannot be greater than MaxRunners '2", "Expected error about MinRunners > MaxRunners")
+}
+
+func TestConfigValidationMissingToken(t *testing.T) {
+	config := &Config{
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+	err := config.validate()
+	expectedError := fmt.Sprintf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
+	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
+}
+
+func TestConfigValidationAppKey(t *testing.T) {
+	config := &Config{
+		AppID:                       1,
+		AppInstallationID:           10,
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+	err := config.validate()
+	expectedError := fmt.Sprintf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
+	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
+}
+
+func TestConfigValidationOnlyOneTypeOfCredentials(t *testing.T) {
+	config := &Config{
+		AppID:                       1,
+		AppInstallationID:           10,
+		AppPrivateKey:               "asdf",
+		Token:                       "asdf",
+		ConfigureUrl:                "github.com/some_org/some_repo",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+	err := config.validate()
+	expectedError := fmt.Sprintf("only one GitHub auth method supported at a time. Have both PAT and App auth: token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
+	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
+}
+
+func TestConfigValidation(t *testing.T) {
+	config := &Config{
+		ConfigureUrl:                "https://github.com/actions",
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+		MinRunners:                  1,
+		MaxRunners:                  5,
+		Token:                       "asdf",
+	}
+
+	err := config.validate()
+
+	assert.NoError(t, err, "Expected no error")
+}
+
+func TestConfigValidationConfigUrl(t *testing.T) {
+	config := &Config{
+		EphemeralRunnerSetNamespace: "namespace",
+		EphemeralRunnerSetName:      "deployment",
+		RunnerScaleSetId:            1,
+	}
+
+	err := config.validate()
+
+	assert.ErrorContains(t, err, "GitHubConfigUrl is not provided", "Expected error about missing ConfigureUrl")
+}

cmd/githubrunnerscalesetlistener/main.go

@@ -28,39 +28,26 @@ import (
 	"time"
 
 	"github.com/actions/actions-runner-controller/build"
+	"github.com/actions/actions-runner-controller/cmd/githubrunnerscalesetlistener/config"
 	"github.com/actions/actions-runner-controller/github/actions"
 	"github.com/actions/actions-runner-controller/logging"
 	"github.com/go-logr/logr"
-	"github.com/kelseyhightower/envconfig"
 	"github.com/prometheus/client_golang/prometheus"
 	"github.com/prometheus/client_golang/prometheus/promhttp"
 	"golang.org/x/net/http/httpproxy"
 	"golang.org/x/sync/errgroup"
 )
 
-type RunnerScaleSetListenerConfig struct {
-	ConfigureUrl                string `split_words:"true"`
-	AppID                       int64  `split_words:"true"`
-	AppInstallationID           int64  `split_words:"true"`
-	AppPrivateKey               string `split_words:"true"`
-	Token                       string `split_words:"true"`
-	EphemeralRunnerSetNamespace string `split_words:"true"`
-	EphemeralRunnerSetName      string `split_words:"true"`
-	MaxRunners                  int    `split_words:"true"`
-	MinRunners                  int    `split_words:"true"`
-	RunnerScaleSetId            int    `split_words:"true"`
-	RunnerScaleSetName          string `split_words:"true"`
-	ServerRootCA                string `split_words:"true"`
-	LogLevel                    string `split_words:"true"`
-	LogFormat                   string `split_words:"true"`
-	MetricsAddr                 string `split_words:"true"`
-	MetricsEndpoint             string `split_words:"true"`
-}
-
 func main() {
-	var rc RunnerScaleSetListenerConfig
-	if err := envconfig.Process("github", &rc); err != nil {
-		fmt.Fprintf(os.Stderr, "Error: processing environment variables for RunnerScaleSetListenerConfig: %v\n", err)
+	configPath, ok := os.LookupEnv("LISTENER_CONFIG_PATH")
+	if !ok {
+		fmt.Fprintf(os.Stderr, "Error: LISTENER_CONFIG_PATH environment variable is not set\n")
+		os.Exit(1)
+	}
+
+	rc, err := config.Read(configPath)
+	if err != nil {
+		fmt.Fprintf(os.Stderr, "Error: reading config from path(%q): %v\n", configPath, err)
 		os.Exit(1)
 	}
 
@@ -80,12 +67,6 @@ func main() {
 		os.Exit(1)
 	}
 
-	// Validate all inputs
-	if err := validateConfig(&rc); err != nil {
-		logger.Error(err, "Inputs validation failed")
-		os.Exit(1)
-	}
-
 	ctx, stop := signal.NotifyContext(context.Background(), syscall.SIGINT, syscall.SIGTERM)
 	defer stop()
 
@@ -123,7 +104,7 @@ func main() {
 }
 
 type metricsServer struct {
-	rc     RunnerScaleSetListenerConfig
+	rc     config.Config
 	logger logr.Logger
 	srv    *http.Server
 }
@@ -173,7 +154,7 @@ type runOptions struct {
 	serviceOptions []func(*Service)
 }
 
-func run(ctx context.Context, rc RunnerScaleSetListenerConfig, logger logr.Logger, opts runOptions) error {
+func run(ctx context.Context, rc config.Config, logger logr.Logger, opts runOptions) error {
 	// Create root context and hook with sigint and sigterm
 	creds := &actions.ActionsAuth{}
 	if rc.Token != "" {
@@ -195,6 +176,8 @@ func run(ctx context.Context, rc RunnerScaleSetListenerConfig, logger logr.Logge
 		Version:    build.Version,
 		CommitSHA:  build.CommitSHA,
 		ScaleSetID: rc.RunnerScaleSetId,
+		HasProxy:   hasProxy(),
+		Subsystem:  "githubrunnerscalesetlistener",
 	})
 	if err != nil {
 		return fmt.Errorf("failed to create an Actions Service client: %w", err)
@@ -232,38 +215,7 @@ func run(ctx context.Context, rc RunnerScaleSetListenerConfig, logger logr.Logge
 	return nil
 }
 
-func validateConfig(config *RunnerScaleSetListenerConfig) error {
-	if len(config.ConfigureUrl) == 0 {
-		return fmt.Errorf("GitHubConfigUrl is not provided")
-	}
-
-	if len(config.EphemeralRunnerSetNamespace) == 0 || len(config.EphemeralRunnerSetName) == 0 {
-		return fmt.Errorf("EphemeralRunnerSetNamespace '%s' or EphemeralRunnerSetName '%s' is missing", config.EphemeralRunnerSetNamespace, config.EphemeralRunnerSetName)
-	}
-
-	if config.RunnerScaleSetId == 0 {
-		return fmt.Errorf("RunnerScaleSetId '%d' is missing", config.RunnerScaleSetId)
-	}
-
-	if config.MaxRunners < config.MinRunners {
-		return fmt.Errorf("MinRunners '%d' cannot be greater than MaxRunners '%d'", config.MinRunners, config.MaxRunners)
-	}
-
-	hasToken := len(config.Token) > 0
-	hasPrivateKeyConfig := config.AppID > 0 && config.AppPrivateKey != ""
-
-	if !hasToken && !hasPrivateKeyConfig {
-		return fmt.Errorf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
-	}
-
-	if hasToken && hasPrivateKeyConfig {
-		return fmt.Errorf("only one GitHub auth method supported at a time. Have both PAT and App auth: token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
-	}
-
-	return nil
-}
-
-func newActionsClientFromConfig(config RunnerScaleSetListenerConfig, creds *actions.ActionsAuth, options ...actions.ClientOption) (*actions.Client, error) {
+func newActionsClientFromConfig(config config.Config, creds *actions.ActionsAuth, options ...actions.ClientOption) (*actions.Client, error) {
 	if config.ServerRootCA != "" {
 		systemPool, err := x509.SystemCertPool()
 		if err != nil {
@@ -285,3 +237,8 @@ func newActionsClientFromConfig(config RunnerScaleSetListenerConfig, creds *acti
 
 	return actions.NewClient(config.ConfigureUrl, creds, options...)
 }
+
+func hasProxy() bool {
+	proxyFunc := httpproxy.FromEnvironment().ProxyFunc()
+	return proxyFunc != nil
+}

cmd/githubrunnerscalesetlistener/main_test.go

@@ -3,7 +3,6 @@ package main
 import (
 	"context"
 	"crypto/tls"
-	"fmt"
 	"net/http"
 	"net/http/httptest"
 	"os"
@@ -13,94 +12,11 @@ import (
 	"github.com/stretchr/testify/assert"
 	"github.com/stretchr/testify/require"
 
+	"github.com/actions/actions-runner-controller/cmd/githubrunnerscalesetlistener/config"
 	"github.com/actions/actions-runner-controller/github/actions"
 	"github.com/actions/actions-runner-controller/github/actions/testserver"
 )
 
-func TestConfigValidationMinMax(t *testing.T) {
-	config := &RunnerScaleSetListenerConfig{
-		ConfigureUrl:                "github.com/some_org/some_repo",
-		EphemeralRunnerSetNamespace: "namespace",
-		EphemeralRunnerSetName:      "deployment",
-		RunnerScaleSetId:            1,
-		MinRunners:                  5,
-		MaxRunners:                  2,
-		Token:                       "token",
-	}
-	err := validateConfig(config)
-	assert.ErrorContains(t, err, "MinRunners '5' cannot be greater than MaxRunners '2", "Expected error about MinRunners > MaxRunners")
-}
-
-func TestConfigValidationMissingToken(t *testing.T) {
-	config := &RunnerScaleSetListenerConfig{
-		ConfigureUrl:                "github.com/some_org/some_repo",
-		EphemeralRunnerSetNamespace: "namespace",
-		EphemeralRunnerSetName:      "deployment",
-		RunnerScaleSetId:            1,
-	}
-	err := validateConfig(config)
-	expectedError := fmt.Sprintf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
-	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
-}
-
-func TestConfigValidationAppKey(t *testing.T) {
-	config := &RunnerScaleSetListenerConfig{
-		AppID:                       1,
-		AppInstallationID:           10,
-		ConfigureUrl:                "github.com/some_org/some_repo",
-		EphemeralRunnerSetNamespace: "namespace",
-		EphemeralRunnerSetName:      "deployment",
-		RunnerScaleSetId:            1,
-	}
-	err := validateConfig(config)
-	expectedError := fmt.Sprintf("GitHub auth credential is missing, token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
-	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
-}
-
-func TestConfigValidationOnlyOneTypeOfCredentials(t *testing.T) {
-	config := &RunnerScaleSetListenerConfig{
-		AppID:                       1,
-		AppInstallationID:           10,
-		AppPrivateKey:               "asdf",
-		Token:                       "asdf",
-		ConfigureUrl:                "github.com/some_org/some_repo",
-		EphemeralRunnerSetNamespace: "namespace",
-		EphemeralRunnerSetName:      "deployment",
-		RunnerScaleSetId:            1,
-	}
-	err := validateConfig(config)
-	expectedError := fmt.Sprintf("only one GitHub auth method supported at a time. Have both PAT and App auth: token length: '%d', appId: '%d', installationId: '%d', private key length: '%d", len(config.Token), config.AppID, config.AppInstallationID, len(config.AppPrivateKey))
-	assert.ErrorContains(t, err, expectedError, "Expected error about missing auth")
-}
-
-func TestConfigValidation(t *testing.T) {
-	config := &RunnerScaleSetListenerConfig{
-		ConfigureUrl:                "https://github.com/actions",
-		EphemeralRunnerSetNamespace: "namespace",
-		EphemeralRunnerSetName:      "deployment",
-		RunnerScaleSetId:            1,
-		MinRunners:                  1,
-		MaxRunners:                  5,
-		Token:                       "asdf",
-	}
-
-	err := validateConfig(config)
-
-	assert.NoError(t, err, "Expected no error")
-}
-
-func TestConfigValidationConfigUrl(t *testing.T) {
-	config := &RunnerScaleSetListenerConfig{
-		EphemeralRunnerSetNamespace: "namespace",
-		EphemeralRunnerSetName:      "deployment",
-		RunnerScaleSetId:            1,
-	}
-
-	err := validateConfig(config)
-
-	assert.ErrorContains(t, err, "GitHubConfigUrl is not provided", "Expected error about missing ConfigureUrl")
-}
-
 func TestCustomerServerRootCA(t *testing.T) {
 	ctx := context.Background()
 	certsFolder := filepath.Join(
@@ -134,7 +50,7 @@ func TestCustomerServerRootCA(t *testing.T) {
 	require.NoError(t, err)
 	certsString = certsString + string(intermediate)
 
-	config := RunnerScaleSetListenerConfig{
+	config := config.Config{
 		ConfigureUrl: server.ConfigURLForOrg("myorg"),
 		ServerRootCA: certsString,
 	}
@@ -164,7 +80,7 @@ func TestProxySettings(t *testing.T) {
 		os.Setenv("http_proxy", proxy.URL)
 		defer os.Setenv("http_proxy", prevProxy)
 
-		config := RunnerScaleSetListenerConfig{
+		config := config.Config{
 			ConfigureUrl: "https://github.com/org/repo",
 		}
 		creds := &actions.ActionsAuth{
@@ -196,7 +112,7 @@ func TestProxySettings(t *testing.T) {
 		os.Setenv("https_proxy", proxy.URL)
 		defer os.Setenv("https_proxy", prevProxy)
 
-		config := RunnerScaleSetListenerConfig{
+		config := config.Config{
 			ConfigureUrl: "https://github.com/org/repo",
 		}
 		creds := &actions.ActionsAuth{
@@ -233,7 +149,7 @@ func TestProxySettings(t *testing.T) {
 		os.Setenv("no_proxy", "example.com")
 		defer os.Setenv("no_proxy", prevNoProxy)
 
-		config := RunnerScaleSetListenerConfig{
+		config := config.Config{
 			ConfigureUrl: "https://github.com/org/repo",
 		}
 		creds := &actions.ActionsAuth{

cmd/githubrunnerscalesetlistener/messageListener.go

@@ -8,6 +8,6 @@ import (
 
 //go:generate mockery --inpackage --name=RunnerScaleSetClient
 type RunnerScaleSetClient interface {
-	GetRunnerScaleSetMessage(ctx context.Context, handler func(msg *actions.RunnerScaleSetMessage) error) error
+	GetRunnerScaleSetMessage(ctx context.Context, handler func(msg *actions.RunnerScaleSetMessage) error, maxCapacity int) error
 	AcquireJobsForRunnerScaleSet(ctx context.Context, requestIds []int64) error
 }

cmd/githubrunnerscalesetlistener/metrics.go

@@ -1,8 +1,6 @@
 package main
 
 import (
-	"strconv"
-
 	"github.com/actions/actions-runner-controller/github/actions"
 	"github.com/prometheus/client_golang/prometheus"
 )
@@ -18,8 +16,6 @@ const (
 	labelKeyJobWorkflowRef          = "job_workflow_ref"
 	labelKeyEventName               = "event_name"
 	labelKeyJobResult               = "job_result"
-	labelKeyRunnerID                = "runner_id"
-	labelKeyRunnerName              = "runner_name"
 )
 
 const githubScaleSetSubsystem = "gha"
@@ -43,10 +39,15 @@ var (
 		labelKeyEventName,
 	}
 
-	completedJobsTotalLabels   = append(jobLabels, labelKeyJobResult, labelKeyRunnerID, labelKeyRunnerName)
-	jobExecutionDurationLabels = append(jobLabels, labelKeyJobResult, labelKeyRunnerID, labelKeyRunnerName)
-	startedJobsTotalLabels     = append(jobLabels, labelKeyRunnerID, labelKeyRunnerName)
-	jobStartupDurationLabels   = append(jobLabels, labelKeyRunnerID, labelKeyRunnerName)
+	completedJobsTotalLabels   = append(jobLabels, labelKeyJobResult)
+	jobExecutionDurationLabels = append(jobLabels, labelKeyJobResult)
+	startedJobsTotalLabels     = jobLabels
+	jobStartupDurationLabels   = []string{
+		labelKeyRepository,
+		labelKeyOrganization,
+		labelKeyEnterprise,
+		labelKeyEventName,
+	}
 )
 
 // metrics
@@ -274,23 +275,34 @@ func (b *baseLabels) scaleSetLabels() prometheus.Labels {
 
 func (b *baseLabels) completedJobLabels(msg *actions.JobCompleted) prometheus.Labels {
 	l := b.jobLabels(&msg.JobMessageBase)
-	l[labelKeyRunnerID] = strconv.Itoa(msg.RunnerId)
 	l[labelKeyJobResult] = msg.Result
-	l[labelKeyRunnerName] = msg.RunnerName
 	return l
 }
 
 func (b *baseLabels) startedJobLabels(msg *actions.JobStarted) prometheus.Labels {
 	l := b.jobLabels(&msg.JobMessageBase)
-	l[labelKeyRunnerID] = strconv.Itoa(msg.RunnerId)
-	l[labelKeyRunnerName] = msg.RunnerName
 	return l
 }
 
+func (b *baseLabels) jobStartupDurationLabels(msg *actions.JobStarted) prometheus.Labels {
+	return prometheus.Labels{
+		labelKeyEnterprise:   b.enterprise,
+		labelKeyOrganization: b.organization,
+		labelKeyRepository:   b.repository,
+		labelKeyEventName:    msg.EventName,
+	}
+}
+
 func (m *metricsExporter) withBaseLabels(base baseLabels) {
 	m.baseLabels = base
 }
 
+func (m *metricsExporter) publishStatic(max, min int) {
+	l := m.scaleSetLabels()
+	maxRunners.With(l).Set(float64(max))
+	minRunners.With(l).Set(float64(min))
+}
+
 func (m *metricsExporter) publishStatistics(stats *actions.RunnerScaleSetStatistic) {
 	l := m.scaleSetLabels()
 
@@ -307,6 +319,7 @@ func (m *metricsExporter) publishJobStarted(msg *actions.JobStarted) {
 	l := m.startedJobLabels(msg)
 	startedJobsTotal.With(l).Inc()
 
+	l = m.jobStartupDurationLabels(msg)
 	startupDuration := msg.JobMessageBase.RunnerAssignTime.Unix() - msg.JobMessageBase.ScaleSetAssignTime.Unix()
 	jobStartupDurationSeconds.With(l).Observe(float64(startupDuration))
 }

cmd/githubrunnerscalesetlistener/mock_KubernetesManager.go

@@ -1,4 +1,4 @@
-// Code generated by mockery v2.33.2. DO NOT EDIT.
+// Code generated by mockery v2.36.1. DO NOT EDIT.
 
 package main
 

cmd/githubrunnerscalesetlistener/mock_RunnerScaleSetClient.go

@@ -1,4 +1,4 @@
-// Code generated by mockery v2.33.2. DO NOT EDIT.
+// Code generated by mockery v2.36.1. DO NOT EDIT.
 
 package main
 
@@ -29,13 +29,13 @@ func (_m *MockRunnerScaleSetClient) AcquireJobsForRunnerScaleSet(ctx context.Con
 	return r0
 }
 
-// GetRunnerScaleSetMessage provides a mock function with given fields: ctx, handler
-func (_m *MockRunnerScaleSetClient) GetRunnerScaleSetMessage(ctx context.Context, handler func(*actions.RunnerScaleSetMessage) error) error {
-	ret := _m.Called(ctx, handler)
+// GetRunnerScaleSetMessage provides a mock function with given fields: ctx, handler, maxCapacity
+func (_m *MockRunnerScaleSetClient) GetRunnerScaleSetMessage(ctx context.Context, handler func(*actions.RunnerScaleSetMessage) error, maxCapacity int) error {
+	ret := _m.Called(ctx, handler, maxCapacity)
 
 	var r0 error
-	if rf, ok := ret.Get(0).(func(context.Context, func(*actions.RunnerScaleSetMessage) error) error); ok {
-		r0 = rf(ctx, handler)
+	if rf, ok := ret.Get(0).(func(context.Context, func(*actions.RunnerScaleSetMessage) error, int) error); ok {
+		r0 = rf(ctx, handler, maxCapacity)
 	} else {
 		r0 = ret.Error(0)
 	}

cmd/githubrunnerscalesetlistener/sessionrefreshingclient.go

@@ -24,8 +24,12 @@ func newSessionClient(client actions.ActionsService, logger *logr.Logger, sessio
 	}
 }
 
-func (m *SessionRefreshingClient) GetMessage(ctx context.Context, lastMessageId int64) (*actions.RunnerScaleSetMessage, error) {
-	message, err := m.client.GetMessage(ctx, m.session.MessageQueueUrl, m.session.MessageQueueAccessToken, lastMessageId)
+func (m *SessionRefreshingClient) GetMessage(ctx context.Context, lastMessageId int64, maxCapacity int) (*actions.RunnerScaleSetMessage, error) {
+	if maxCapacity < 0 {
+		return nil, fmt.Errorf("maxCapacity must be greater than or equal to 0")
+	}
+
+	message, err := m.client.GetMessage(ctx, m.session.MessageQueueUrl, m.session.MessageQueueAccessToken, lastMessageId, maxCapacity)
 	if err == nil {
 		return message, nil
 	}
@@ -42,7 +46,7 @@ func (m *SessionRefreshingClient) GetMessage(ctx context.Context, lastMessageId
 	}
 
 	m.session = session
-	message, err = m.client.GetMessage(ctx, m.session.MessageQueueUrl, m.session.MessageQueueAccessToken, lastMessageId)
+	message, err = m.client.GetMessage(ctx, m.session.MessageQueueUrl, m.session.MessageQueueAccessToken, lastMessageId, maxCapacity)
 	if err != nil {
 		return nil, fmt.Errorf("delete message failed after refresh message session. %w", err)
 	}

cmd/githubrunnerscalesetlistener/sessionrefreshingclient_test.go

@@ -31,17 +31,17 @@ func TestGetMessage(t *testing.T) {
 		},
 	}
 
-	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0)).Return(nil, nil).Once()
-	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0)).Return(&actions.RunnerScaleSetMessage{MessageId: 1}, nil).Once()
+	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0), 10).Return(nil, nil).Once()
+	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0), 10).Return(&actions.RunnerScaleSetMessage{MessageId: 1}, nil).Once()
 
 	client := newSessionClient(mockActionsClient, &logger, session)
 
-	msg, err := client.GetMessage(ctx, 0)
+	msg, err := client.GetMessage(ctx, 0, 10)
 	require.NoError(t, err, "GetMessage should not return an error")
 
 	assert.Nil(t, msg, "GetMessage should return nil message")
 
-	msg, err = client.GetMessage(ctx, 0)
+	msg, err = client.GetMessage(ctx, 0, 10)
 	require.NoError(t, err, "GetMessage should not return an error")
 
 	assert.Equal(t, int64(1), msg.MessageId, "GetMessage should return a message with id 1")
@@ -146,11 +146,11 @@ func TestGetMessage_Error(t *testing.T) {
 		},
 	}
 
-	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0)).Return(nil, fmt.Errorf("error")).Once()
+	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0), 10).Return(nil, fmt.Errorf("error")).Once()
 
 	client := newSessionClient(mockActionsClient, &logger, session)
 
-	msg, err := client.GetMessage(ctx, 0)
+	msg, err := client.GetMessage(ctx, 0, 10)
 	assert.ErrorContains(t, err, "get message failed. error", "GetMessage should return an error")
 	assert.Nil(t, msg, "GetMessage should return nil message")
 	assert.True(t, mockActionsClient.AssertExpectations(t), "All expected calls to mockActionsClient should have been made")
@@ -227,8 +227,8 @@ func TestGetMessage_RefreshToken(t *testing.T) {
 			Id: 1,
 		},
 	}
-	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0)).Return(nil, &actions.MessageQueueTokenExpiredError{}).Once()
-	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, "token2", int64(0)).Return(&actions.RunnerScaleSetMessage{
+	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0), 10).Return(nil, &actions.MessageQueueTokenExpiredError{}).Once()
+	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, "token2", int64(0), 10).Return(&actions.RunnerScaleSetMessage{
 		MessageId:   1,
 		MessageType: "test",
 		Body:        "test",
@@ -243,7 +243,7 @@ func TestGetMessage_RefreshToken(t *testing.T) {
 	}, nil).Once()
 
 	client := newSessionClient(mockActionsClient, &logger, session)
-	msg, err := client.GetMessage(ctx, 0)
+	msg, err := client.GetMessage(ctx, 0, 10)
 	assert.NoError(t, err, "Error getting message")
 	assert.Equal(t, int64(1), msg.MessageId, "message id should be updated")
 	assert.Equal(t, "token2", client.session.MessageQueueAccessToken, "Message queue access token should be updated")
@@ -340,11 +340,11 @@ func TestGetMessage_RefreshToken_Failed(t *testing.T) {
 			Id: 1,
 		},
 	}
-	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0)).Return(nil, &actions.MessageQueueTokenExpiredError{}).Once()
+	mockActionsClient.On("GetMessage", ctx, session.MessageQueueUrl, session.MessageQueueAccessToken, int64(0), 10).Return(nil, &actions.MessageQueueTokenExpiredError{}).Once()
 	mockActionsClient.On("RefreshMessageSession", ctx, session.RunnerScaleSet.Id, session.SessionId).Return(nil, fmt.Errorf("error"))
 
 	client := newSessionClient(mockActionsClient, &logger, session)
-	msg, err := client.GetMessage(ctx, 0)
+	msg, err := client.GetMessage(ctx, 0, 10)
 	assert.ErrorContains(t, err, "refresh message session failed. error", "Error should be returned")
 	assert.Nil(t, msg, "Message should be nil")
 	assert.Equal(t, "token", client.session.MessageQueueAccessToken, "Message queue access token should not be updated")

cmd/githubwebhookserver/main.go

@@ -39,6 +39,9 @@ import (
 	_ "k8s.io/client-go/plugin/pkg/client/auth/gcp"
 	_ "k8s.io/client-go/plugin/pkg/client/auth/oidc"
 	ctrl "sigs.k8s.io/controller-runtime"
+	"sigs.k8s.io/controller-runtime/pkg/cache"
+	metricsserver "sigs.k8s.io/controller-runtime/pkg/metrics/server"
+	"sigs.k8s.io/controller-runtime/pkg/webhook"
 	// +kubebuilder:scaffold:imports
 )
 
@@ -150,10 +153,18 @@ func main() {
 	syncPeriod := 10 * time.Minute
 	mgr, err := ctrl.NewManager(ctrl.GetConfigOrDie(), ctrl.Options{
 		Scheme: scheme,
+		Cache: cache.Options{
 			SyncPeriod: &syncPeriod,
-		Namespace:          watchNamespace,
-		MetricsBindAddress: metricsAddr,
+			DefaultNamespaces: map[string]cache.Config{
+				watchNamespace: {},
+			},
+		},
+		Metrics: metricsserver.Options{
+			BindAddress: metricsAddr,
+		},
+		WebhookServer: webhook.NewServer(webhook.Options{
 			Port: 9443,
+		}),
 	})
 	if err != nil {
 		logger.Error(err, "unable to start manager")

config/crd/bases/actions.summerwind.dev_horizontalrunnerautoscalers.yaml

@@ -3,8 +3,7 @@ apiVersion: apiextensions.k8s.io/v1
 kind: CustomResourceDefinition
 metadata:
   annotations:
-    controller-gen.kubebuilder.io/version: v0.11.3
-  creationTimestamp: null
+    controller-gen.kubebuilder.io/version: v0.14.0
   name: horizontalrunnerautoscalers.actions.summerwind.dev
 spec:
   group: actions.summerwind.dev
@@ -36,10 +35,19 @@ spec:
           description: HorizontalRunnerAutoscaler is the Schema for the horizontalrunnerautoscaler API
           properties:
             apiVersion:
-              description: 'APIVersion defines the versioned schema of this representation of an object. Servers should convert recognized schemas to the latest internal value, and may reject unrecognized values. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources'
+              description: |-
+                APIVersion defines the versioned schema of this representation of an object.
+                Servers should convert recognized schemas to the latest internal value, and
+                may reject unrecognized values.
+                More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#resources
               type: string
             kind:
-              description: 'Kind is a string value representing the REST resource this object represents. Servers may infer this from the endpoint the client submits requests to. Cannot be updated. In CamelCase. More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds'
+              description: |-
+                Kind is a string value representing the REST resource this object represents.
+                Servers may infer this from the endpoint the client submits requests to.
+                Cannot be updated.
+                In CamelCase.
+                More info: https://git.k8s.io/community/contributors/devel/sig-architecture/api-conventions.md#types-kinds
               type: string
             metadata:
               type: object
@@ -48,7 +56,9 @@ spec:
               properties:
                 capacityReservations:
                   items:
-                    description: CapacityReservation specifies the number of replicas temporarily added to the scale target until ExpirationTime.
+                    description: |-
+                      CapacityReservation specifies the number of replicas temporarily added
+                      to the scale target until ExpirationTime.
                     properties:
                       effectiveTime:
                         format: date-time
@@ -80,30 +90,46 @@ spec:
                   items:
                     properties:
                       repositoryNames:
-                        description: RepositoryNames is the list of repository names to be used for calculating the metric. For example, a repository name is the REPO part of `github.com/USER/REPO`.
+                        description: |-
+                          RepositoryNames is the list of repository names to be used for calculating the metric.
+                          For example, a repository name is the REPO part of `github.com/USER/REPO`.
                         items:
                           type: string
                         type: array
                       scaleDownAdjustment:
-                        description: ScaleDownAdjustment is the number of runners removed on scale-down. You can only specify either ScaleDownFactor or ScaleDownAdjustment.
+                        description: |-
+                          ScaleDownAdjustment is the number of runners removed on scale-down.
+                          You can only specify either ScaleDownFactor or ScaleDownAdjustment.
                         type: integer
                       scaleDownFactor:
-                        description: ScaleDownFactor is the multiplicative factor applied to the current number of runners used to determine how many pods should be removed.
+                        description: |-
+                          ScaleDownFactor is the multiplicative factor applied to the current number of runners used
+                          to determine how many pods should be removed.
                         type: string
                       scaleDownThreshold:
-                        description: ScaleDownThreshold is the percentage of busy runners less than which will trigger the hpa to scale the runners down.
+                        description: |-
+                          ScaleDownThreshold is the percentage of busy runners less than which will
+                          trigger the hpa to scale the runners down.
                         type: string
                       scaleUpAdjustment:
-                        description: ScaleUpAdjustment is the number of runners added on scale-up. You can only specify either ScaleUpFactor or ScaleUpAdjustment.
+                        description: |-
+                          ScaleUpAdjustment is the number of runners added on scale-up.
+                          You can only specify either ScaleUpFactor or ScaleUpAdjustment.
                         type: integer
                       scaleUpFactor:
-                        description: ScaleUpFactor is the multiplicative factor applied to the current number of runners used to determine how many pods should be added.
+                        description: |-
+                          ScaleUpFactor is the multiplicative factor applied to the current number of runners used
+                          to determine how many pods should be added.
                         type: string
                       scaleUpThreshold:
-                        description: ScaleUpThreshold is the percentage of busy runners greater than which will trigger the hpa to scale runners up.
+                        description: |-
+                          ScaleUpThreshold is the percentage of busy runners greater than which will
+                          trigger the hpa to scale runners up.
                         type: string
                       type:
-                        description: Type is the type of metric to be used for autoscaling. It can be TotalNumberOfQueuedAndInProgressWorkflowRuns or PercentageRunnersBusy.
+                        description: |-
+                          Type is the type of metric to be used for autoscaling.
+                          It can be TotalNumberOfQueuedAndInProgressWorkflowRuns or PercentageRunnersBusy.
                         type: string
                     type: object
                   type: array
@@ -111,7 +137,9 @@ spec:
                   description: MinReplicas is the minimum number of replicas the deployment is allowed to scale
                   type: integer
                 scaleDownDelaySecondsAfterScaleOut:
-                  description: ScaleDownDelaySecondsAfterScaleUp is the approximate delay for a scale down followed by a scale up Used to prevent flapping (down->up->down->... loop)
+                  description: |-
+                    ScaleDownDelaySecondsAfterScaleUp is the approximate delay for a scale down followed by a scale up
+                    Used to prevent flapping (down->up->down->... loop)
                   type: integer
                 scaleTargetRef:
                   description: ScaleTargetRef is the reference to scaled resource like RunnerDeployment
@@ -127,7 +155,18 @@ spec:
                       type: string
                   type: object
                 scaleUpTriggers:
-                  description: "ScaleUpTriggers is an experimental feature to increase the desired replicas by 1 on each webhook requested received by the webhookBasedAutoscaler. \n This feature requires you to also enable and deploy the webhookBasedAutoscaler onto your cluster. \n Note that the added runners remain until the next sync period at least, and they may or may not be used by GitHub Actions depending on the timing. They are intended to be used to gain \"resource slack\" immediately after you receive a webhook from GitHub, so that you can loosely expect MinReplicas runners to be always available."
+                  description: |-
+                    ScaleUpTriggers is an experimental feature to increase the desired replicas by 1
+                    on each webhook requested received by the webhookBasedAutoscaler.
+
+
+                    This feature requires you to also enable and deploy the webhookBasedAutoscaler onto your cluster.
+
+
+                    Note that the added runners remain until the next sync period at least,
+                    and they may or may not be used by GitHub Actions depending on the timing.
+                    They are intended to be used to gain "resource slack" immediately after you
+                    receive a webhook from GitHub, so that you can loosely expect MinReplicas runners to be always available.
                   items:
                     properties:
                       amount:
@@ -140,12 +179,18 @@ spec:
                             description: https://docs.github.com/en/actions/reference/events-that-trigger-workflows#check_run
                             properties:
                               names:
-                                description: Names is a list of GitHub Actions glob patterns. Any check_run event whose name matches one of patterns in the list can trigger autoscaling. Note that check_run name seem to equal to the job name you've defined in your actions workflow yaml file. So it is very likely that you can utilize this to trigger depending on the job.
+                                description: |-
+                                  Names is a list of GitHub Actions glob patterns.
+                                  Any check_run event whose name matches one of patterns in the list can trigger autoscaling.
+                                  Note that check_run name seem to equal to the job name you've defined in your actions workflow yaml file.
+                                  So it is very likely that you can utilize this to trigger depending on the job.
                                 items:
                                   type: string
                                 type: array
                               repositories:
-                                description: Repositories is a list of GitHub repositories. Any check_run event whose repository matches one of repositories in the list can trigger autoscaling.
+                                description: |-
+                                  Repositories is a list of GitHub repositories.
+                                  Any check_run event whose repository matches one of repositories in the list can trigger autoscaling.
                                 items:
                                   type: string
                                 type: array
@@ -170,7 +215,9 @@ spec:
                                 type: array
                             type: object
                           push:
-                            description: PushSpec is the condition for triggering scale-up on push event Also see https://docs.github.com/en/actions/reference/events-that-trigger-workflows#push
+                            description: |-
+                              PushSpec is the condition for triggering scale-up on push event
+                              Also see https://docs.github.com/en/actions/reference/events-that-trigger-workflows#push
                             type: object
                           workflowJob:
                             description: https://docs.github.com/en/developers/webhooks-and-events/webhooks/webhook-events-and-payloads#workflow_job
@@ -179,23 +226,33 @@ spec:
                     type: object
                   type: array
                 scheduledOverrides:
-                  description: ScheduledOverrides is the list of ScheduledOverride. It can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule. The earlier a scheduled override is, the higher it is prioritized.
+                  description: |-
+                    ScheduledOverrides is the list of ScheduledOverride.
+                    It can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule.
+                    The earlier a scheduled override is, the higher it is prioritized.
                   items:
-                    description: ScheduledOverride can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule. A schedule can optionally be recurring, so that the corresponding override happens every day, week, month, or year.
+                    description: |-
+                      ScheduledOverride can be used to override a few fields of HorizontalRunnerAutoscalerSpec on schedule.
+                      A schedule can optionally be recurring, so that the corresponding override happens every day, week, month, or year.
                     properties:
                       endTime:
                         description: EndTime is the time at which the first override ends.
                         format: date-time
                         type: string
                       minReplicas:
-                        description: MinReplicas is the number of runners while overriding. If omitted, it doesn't override minReplicas.
+                        description: |-
+                          MinReplicas is the number of runners while overriding.
+                          If omitted, it doesn't override minReplicas.
                         minimum: 0
                         nullable: true
                         type: integer
                       recurrenceRule:
                         properties:
                           frequency:
-                            description: Frequency is the name of a predefined interval of each recurrence. The valid values are "Daily", "Weekly", "Monthly", and "Yearly". If empty, the corresponding override happens only once.
+                            description: |-
+                              Frequency is the name of a predefined interval of each recurrence.
+                              The valid values are "Daily", "Weekly", "Monthly", and "Yearly".
+                              If empty, the corresponding override happens only once.
                             enum:
                               - Daily
                               - Weekly
@@ -203,7 +260,9 @@ spec:
                               - Yearly
                             type: string
                           untilTime:
-                            description: UntilTime is the time of the final recurrence. If empty, the schedule recurs forever.
+                            description: |-
+                              UntilTime is the time of the final recurrence.
+                              If empty, the schedule recurs forever.
                             format: date-time
                             type: string
                         type: object
@@ -232,18 +291,24 @@ spec:
                     type: object
                   type: array
                 desiredReplicas:
-                  description: DesiredReplicas is the total number of desired, non-terminated and latest pods to be set for the primary RunnerSet This doesn't include outdated pods while upgrading the deployment and replacing the runnerset.
+                  description: |-
+                    DesiredReplicas is the total number of desired, non-terminated and latest pods to be set for the primary RunnerSet
+                    This doesn't include outdated pods while upgrading the deployment and replacing the runnerset.
                   type: integer
                 lastSuccessfulScaleOutTime:
                   format: date-time
                   nullable: true
                   type: string
                 observedGeneration:
-                  description: ObservedGeneration is the most recent generation observed for the target. It corresponds to e.g. RunnerDeployment's generation, which is updated on mutation by the API Server.
+                  description: |-
+                    ObservedGeneration is the most recent generation observed for the target. It corresponds to e.g.
+                    RunnerDeployment's generation, which is updated on mutation by the API Server.
                   format: int64
                   type: integer
                 scheduledOverridesSummary:
-                  description: ScheduledOverridesSummary is the summary of active and upcoming scheduled overrides to be shown in e.g. a column of a `kubectl get hra` output for observability.
+                  description: |-
+                    ScheduledOverridesSummary is the summary of active and upcoming scheduled overrides to be shown in e.g. a column of a `kubectl get hra` output
+                    for observability.
                   type: string
               type: object
           type: object

config/rbac/role.yaml

@@ -2,7 +2,6 @@
 apiVersion: rbac.authorization.k8s.io/v1
 kind: ClusterRole
 metadata:
-  creationTimestamp: null
   name: manager-role
 rules:
 - apiGroups:

config/webhook/manifests.yaml

@@ -2,7 +2,6 @@
 apiVersion: admissionregistration.k8s.io/v1
 kind: MutatingWebhookConfiguration
 metadata:
-  creationTimestamp: null
   name: mutating-webhook-configuration
 webhooks:
 - admissionReviewVersions:
@@ -88,7 +87,6 @@ webhooks:
 apiVersion: admissionregistration.k8s.io/v1
 kind: ValidatingWebhookConfiguration
 metadata:
-  creationTimestamp: null
   name: validating-webhook-configuration
 webhooks:
 - admissionReviewVersions:

controllers/actions.github.com/autoscalinglistener_controller.go

@@ -30,7 +30,6 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/handler"
 	"sigs.k8s.io/controller-runtime/pkg/predicate"
 	"sigs.k8s.io/controller-runtime/pkg/reconcile"
-	"sigs.k8s.io/controller-runtime/pkg/source"
 
 	v1alpha1 "github.com/actions/actions-runner-controller/apis/actions.github.com/v1alpha1"
 	"github.com/actions/actions-runner-controller/controllers/actions.github.com/metrics"
@@ -42,7 +41,7 @@ import (
 )
 
 const (
-	autoscalingListenerContainerName = "autoscaler"
+	autoscalingListenerContainerName = "listener"
 	autoscalingListenerFinalizerName = "autoscalinglistener.actions.github.com/finalizer"
 )
 
@@ -287,6 +286,21 @@ func (r *AutoscalingListenerReconciler) cleanupResources(ctx context.Context, au
 	}
 	logger.Info("Listener pod is deleted")
 
+	var secret corev1.Secret
+	err = r.Get(ctx, types.NamespacedName{Namespace: autoscalingListener.Namespace, Name: scaleSetListenerConfigName(autoscalingListener)}, &secret)
+	switch {
+	case err == nil:
+		if secret.ObjectMeta.DeletionTimestamp.IsZero() {
+			logger.Info("Deleting the listener config secret")
+			if err := r.Delete(ctx, &secret); err != nil {
+				return false, fmt.Errorf("failed to delete listener config secret: %v", err)
+			}
+		}
+		return false, nil
+	case err != nil && !kerrors.IsNotFound(err):
+		return false, fmt.Errorf("failed to get listener config secret: %v", err)
+	}
+
 	if autoscalingListener.Spec.Proxy != nil {
 		logger.Info("Cleaning up the listener proxy secret")
 		proxySecret := new(corev1.Secret)
@@ -306,6 +320,38 @@ func (r *AutoscalingListenerReconciler) cleanupResources(ctx context.Context, au
 		logger.Info("Listener proxy secret is deleted")
 	}
 
+	listenerRoleBinding := new(rbacv1.RoleBinding)
+	err = r.Get(ctx, types.NamespacedName{Namespace: autoscalingListener.Spec.AutoscalingRunnerSetNamespace, Name: scaleSetListenerRoleName(autoscalingListener)}, listenerRoleBinding)
+	switch {
+	case err == nil:
+		if listenerRoleBinding.ObjectMeta.DeletionTimestamp.IsZero() {
+			logger.Info("Deleting the listener role binding")
+			if err := r.Delete(ctx, listenerRoleBinding); err != nil {
+				return false, fmt.Errorf("failed to delete listener role binding: %v", err)
+			}
+		}
+		return false, nil
+	case err != nil && !kerrors.IsNotFound(err):
+		return false, fmt.Errorf("failed to get listener role binding: %v", err)
+	}
+	logger.Info("Listener role binding is deleted")
+
+	listenerRole := new(rbacv1.Role)
+	err = r.Get(ctx, types.NamespacedName{Namespace: autoscalingListener.Spec.AutoscalingRunnerSetNamespace, Name: scaleSetListenerRoleName(autoscalingListener)}, listenerRole)
+	switch {
+	case err == nil:
+		if listenerRole.ObjectMeta.DeletionTimestamp.IsZero() {
+			logger.Info("Deleting the listener role")
+			if err := r.Delete(ctx, listenerRole); err != nil {
+				return false, fmt.Errorf("failed to delete listener role: %v", err)
+			}
+		}
+		return false, nil
+	case err != nil && !kerrors.IsNotFound(err):
+		return false, fmt.Errorf("failed to get listener role: %v", err)
+	}
+	logger.Info("Listener role is deleted")
+
 	logger.Info("Cleaning up the listener service account")
 	listenerSa := new(corev1.ServiceAccount)
 	err = r.Get(ctx, types.NamespacedName{Name: scaleSetListenerServiceAccountName(autoscalingListener), Namespace: autoscalingListener.Namespace}, listenerSa)
@@ -386,13 +432,13 @@ func (r *AutoscalingListenerReconciler) createListenerPod(ctx context.Context, a
 		}
 	}
 
+	cert := ""
 	if autoscalingListener.Spec.GitHubServerTLS != nil {
-		env, err := r.certificateEnvVarForListener(ctx, autoscalingRunnerSet, autoscalingListener)
+		var err error
+		cert, err = r.certificate(ctx, autoscalingRunnerSet, autoscalingListener)
 		if err != nil {
 			return ctrl.Result{}, fmt.Errorf("failed to create certificate env var for listener: %v", err)
 		}
-
-		envs = append(envs, env)
 	}
 
 	var metricsConfig *listenerMetricsServerConfig
@@ -403,7 +449,35 @@ func (r *AutoscalingListenerReconciler) createListenerPod(ctx context.Context, a
 		}
 	}
 
-	newPod, err := r.resourceBuilder.newScaleSetListenerPod(autoscalingListener, serviceAccount, secret, metricsConfig, envs...)
+	var podConfig corev1.Secret
+	if err := r.Get(ctx, types.NamespacedName{Namespace: autoscalingListener.Namespace, Name: scaleSetListenerConfigName(autoscalingListener)}, &podConfig); err != nil {
+		if !kerrors.IsNotFound(err) {
+			logger.Error(err, "Unable to get listener config secret", "namespace", autoscalingListener.Namespace, "name", scaleSetListenerConfigName(autoscalingListener))
+			return ctrl.Result{Requeue: true}, err
+		}
+
+		logger.Info("Creating listener config secret")
+
+		podConfig, err := r.resourceBuilder.newScaleSetListenerConfig(autoscalingListener, secret, metricsConfig, cert)
+		if err != nil {
+			logger.Error(err, "Failed to build listener config secret")
+			return ctrl.Result{}, err
+		}
+
+		if err := ctrl.SetControllerReference(autoscalingListener, podConfig, r.Scheme); err != nil {
+			logger.Error(err, "Failed to set controller reference")
+			return ctrl.Result{}, err
+		}
+
+		if err := r.Create(ctx, podConfig); err != nil {
+			logger.Error(err, "Unable to create listener config secret", "namespace", podConfig.Namespace, "name", podConfig.Name)
+			return ctrl.Result{}, err
+		}
+
+		return ctrl.Result{Requeue: true}, nil
+	}
+
+	newPod, err := r.resourceBuilder.newScaleSetListenerPod(autoscalingListener, &podConfig, serviceAccount, secret, metricsConfig, envs...)
 	if err != nil {
 		logger.Error(err, "Failed to build listener pod")
 		return ctrl.Result{}, err
@@ -424,13 +498,13 @@ func (r *AutoscalingListenerReconciler) createListenerPod(ctx context.Context, a
 	return ctrl.Result{}, nil
 }
 
-func (r *AutoscalingListenerReconciler) certificateEnvVarForListener(ctx context.Context, autoscalingRunnerSet *v1alpha1.AutoscalingRunnerSet, autoscalingListener *v1alpha1.AutoscalingListener) (corev1.EnvVar, error) {
+func (r *AutoscalingListenerReconciler) certificate(ctx context.Context, autoscalingRunnerSet *v1alpha1.AutoscalingRunnerSet, autoscalingListener *v1alpha1.AutoscalingListener) (string, error) {
 	if autoscalingListener.Spec.GitHubServerTLS.CertificateFrom == nil {
-		return corev1.EnvVar{}, fmt.Errorf("githubServerTLS.certificateFrom is not specified")
+		return "", fmt.Errorf("githubServerTLS.certificateFrom is not specified")
 	}
 
 	if autoscalingListener.Spec.GitHubServerTLS.CertificateFrom.ConfigMapKeyRef == nil {
-		return corev1.EnvVar{}, fmt.Errorf("githubServerTLS.certificateFrom.configMapKeyRef is not specified")
+		return "", fmt.Errorf("githubServerTLS.certificateFrom.configMapKeyRef is not specified")
 	}
 
 	var configmap corev1.ConfigMap
@@ -443,7 +517,7 @@ func (r *AutoscalingListenerReconciler) certificateEnvVarForListener(ctx context
 		&configmap,
 	)
 	if err != nil {
-		return corev1.EnvVar{}, fmt.Errorf(
+		return "", fmt.Errorf(
 			"failed to get configmap %s: %w",
 			autoscalingListener.Spec.GitHubServerTLS.CertificateFrom.ConfigMapKeyRef.Name,
 			err,
@@ -452,17 +526,14 @@ func (r *AutoscalingListenerReconciler) certificateEnvVarForListener(ctx context
 
 	certificate, ok := configmap.Data[autoscalingListener.Spec.GitHubServerTLS.CertificateFrom.ConfigMapKeyRef.Key]
 	if !ok {
-		return corev1.EnvVar{}, fmt.Errorf(
+		return "", fmt.Errorf(
 			"key %s is not found in configmap %s",
 			autoscalingListener.Spec.GitHubServerTLS.CertificateFrom.ConfigMapKeyRef.Key,
 			autoscalingListener.Spec.GitHubServerTLS.CertificateFrom.ConfigMapKeyRef.Name,
 		)
 	}
 
-	return corev1.EnvVar{
-		Name:  "GITHUB_SERVER_ROOT_CA",
-		Value: certificate,
-	}, nil
+	return certificate, nil
 }
 
 func (r *AutoscalingListenerReconciler) createSecretsForListener(ctx context.Context, autoscalingListener *v1alpha1.AutoscalingListener, secret *corev1.Secret, logger logr.Logger) (ctrl.Result, error) {
@@ -643,7 +714,7 @@ func (r *AutoscalingListenerReconciler) SetupWithManager(mgr ctrl.Manager) error
 		return err
 	}
 
-	labelBasedWatchFunc := func(obj client.Object) []reconcile.Request {
+	labelBasedWatchFunc := func(_ context.Context, obj client.Object) []reconcile.Request {
 		var requests []reconcile.Request
 		labels := obj.GetLabels()
 		namespace, ok := labels["auto-scaling-listener-namespace"]
@@ -670,8 +741,8 @@ func (r *AutoscalingListenerReconciler) SetupWithManager(mgr ctrl.Manager) error
 		For(&v1alpha1.AutoscalingListener{}).
 		Owns(&corev1.Pod{}).
 		Owns(&corev1.ServiceAccount{}).
-		Watches(&source.Kind{Type: &rbacv1.Role{}}, handler.EnqueueRequestsFromMapFunc(labelBasedWatchFunc)).
-		Watches(&source.Kind{Type: &rbacv1.RoleBinding{}}, handler.EnqueueRequestsFromMapFunc(labelBasedWatchFunc)).
+		Watches(&rbacv1.Role{}, handler.EnqueueRequestsFromMapFunc(labelBasedWatchFunc)).
+		Watches(&rbacv1.RoleBinding{}, handler.EnqueueRequestsFromMapFunc(labelBasedWatchFunc)).
 		WithEventFilter(predicate.ResourceVersionChangedPredicate{}).
 		Complete(r)
 }

controllers/actions.github.com/autoscalinglistener_controller_test.go

@@ -2,6 +2,7 @@ package actionsgithubcom
 
 import (
 	"context"
+	"encoding/json"
 	"fmt"
 	"os"
 	"path/filepath"
@@ -13,6 +14,7 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/client"
 	logf "sigs.k8s.io/controller-runtime/pkg/log"
 
+	listenerconfig "github.com/actions/actions-runner-controller/cmd/githubrunnerscalesetlistener/config"
 	. "github.com/onsi/ginkgo/v2"
 	. "github.com/onsi/gomega"
 	kerrors "k8s.io/apimachinery/pkg/api/errors"
@@ -103,6 +105,19 @@ var _ = Describe("Test AutoScalingListener controller", func() {
 
 	Context("When creating a new AutoScalingListener", func() {
 		It("It should create/add all required resources for a new AutoScalingListener (finalizer, secret, service account, role, rolebinding, pod)", func() {
+			config := new(corev1.Secret)
+			Eventually(
+				func() error {
+					err := k8sClient.Get(ctx, client.ObjectKey{Name: scaleSetListenerConfigName(autoscalingListener), Namespace: configSecret.Namespace}, config)
+					if err != nil {
+						return err
+					}
+					return nil
+				},
+				autoscalingListenerTestTimeout,
+				autoscalingListenerTestInterval,
+			).Should(Succeed(), "Config secret should be created")
+
 			// Check if finalizer is added
 			created := new(actionsv1alpha1.AutoscalingListener)
 			Eventually(
@@ -203,7 +218,8 @@ var _ = Describe("Test AutoScalingListener controller", func() {
 					return pod.Name, nil
 				},
 				autoscalingListenerTestTimeout,
-				autoscalingListenerTestInterval).Should(BeEquivalentTo(autoscalingListener.Name), "Pod should be created")
+				autoscalingListenerTestInterval,
+			).Should(BeEquivalentTo(autoscalingListener.Name), "Pod should be created")
 
 			// Delete the AutoScalingListener
 			err := k8sClient.Delete(ctx, autoscalingListener)
@@ -225,7 +241,41 @@ var _ = Describe("Test AutoScalingListener controller", func() {
 					return nil
 				},
 				autoscalingListenerTestTimeout,
-				autoscalingListenerTestInterval).ShouldNot(Succeed(), "failed to delete pod")
+				autoscalingListenerTestInterval,
+			).ShouldNot(Succeed(), "failed to delete pod")
+
+			// Cleanup the listener role binding
+			Eventually(
+				func() bool {
+					roleBinding := new(rbacv1.RoleBinding)
+					err := k8sClient.Get(ctx, client.ObjectKey{Name: scaleSetListenerRoleName(autoscalingListener), Namespace: autoscalingListener.Spec.AutoscalingRunnerSetNamespace}, roleBinding)
+					return kerrors.IsNotFound(err)
+				},
+				autoscalingListenerTestTimeout,
+				autoscalingListenerTestInterval,
+			).Should(BeTrue(), "failed to delete role binding")
+
+			// Cleanup the listener role
+			Eventually(
+				func() bool {
+					role := new(rbacv1.Role)
+					err := k8sClient.Get(ctx, client.ObjectKey{Name: scaleSetListenerRoleName(autoscalingListener), Namespace: autoscalingListener.Spec.AutoscalingRunnerSetNamespace}, role)
+					return kerrors.IsNotFound(err)
+				},
+				autoscalingListenerTestTimeout,
+				autoscalingListenerTestInterval,
+			).Should(BeTrue(), "failed to delete role")
+
+			// Cleanup the listener config
+			Eventually(
+				func() bool {
+					config := new(corev1.Secret)
+					err := k8sClient.Get(ctx, client.ObjectKey{Name: scaleSetListenerConfigName(autoscalingListener), Namespace: autoscalingListener.Namespace}, config)
+					return kerrors.IsNotFound(err)
+				},
+				autoscalingListenerTestTimeout,
+				autoscalingListenerTestInterval,
+			).Should(BeTrue(), "failed to delete config secret")
 
 			// Cleanup the listener service account
 			Eventually(
@@ -375,7 +425,7 @@ var _ = Describe("Test AutoScalingListener customization", func() {
 		Spec: corev1.PodSpec{
 			Containers: []corev1.Container{
 				{
-					Name:            "listener",
+					Name:            autoscalingListenerContainerName,
 					ImagePullPolicy: corev1.PullAlways,
 					SecurityContext: &corev1.SecurityContext{
 						RunAsUser: &runAsUser,
@@ -477,6 +527,17 @@ var _ = Describe("Test AutoScalingListener customization", func() {
 				autoscalingListenerTestTimeout,
 				autoscalingListenerTestInterval).Should(BeEquivalentTo(autoscalingListenerFinalizerName), "AutoScalingListener should have a finalizer")
 
+			// Check if config is created
+			config := new(corev1.Secret)
+			Eventually(
+				func() error {
+					err := k8sClient.Get(ctx, client.ObjectKey{Name: scaleSetListenerConfigName(autoscalingListener), Namespace: autoscalingListener.Namespace}, config)
+					return err
+				},
+				autoscalingListenerTestTimeout,
+				autoscalingListenerTestInterval,
+			).Should(Succeed(), "Config secret should be created")
+
 			// Check if pod is created
 			pod := new(corev1.Pod)
 			Eventually(
@@ -494,7 +555,7 @@ var _ = Describe("Test AutoScalingListener customization", func() {
 
 			Expect(pod.Spec.SecurityContext.RunAsUser).To(Equal(&runAsUser), "Pod should have the correct security context")
 
-			Expect(pod.Spec.Containers[0].Name).NotTo(Equal("listener"), "Pod should have the correct container name")
+			Expect(pod.Spec.Containers[0].Name).To(Equal(autoscalingListenerContainerName), "Pod should have the correct container name")
 			Expect(pod.Spec.Containers[0].SecurityContext.RunAsUser).To(Equal(&runAsUser), "Pod should have the correct security context")
 			Expect(pod.Spec.Containers[0].ImagePullPolicy).To(Equal(corev1.PullAlways), "Pod should have the correct image pull policy")
 
@@ -793,7 +854,7 @@ var _ = Describe("Test AutoScalingListener controller with template modification
 			Spec: corev1.PodSpec{
 				Containers: []corev1.Container{
 					{
-						Name:            "listener",
+						Name:            autoscalingListenerContainerName,
 						ImagePullPolicy: corev1.PullAlways,
 						SecurityContext: &corev1.SecurityContext{
 							RunAsUser: &runAsUser1001,
@@ -965,31 +1026,26 @@ var _ = Describe("Test GitHub Server TLS configuration", func() {
 	})
 
 	Context("When creating a new AutoScalingListener", func() {
-		It("It should set the certificates as an environment variable on the pod", func() {
-			pod := new(corev1.Pod)
+		It("It should set the certificates in the config of the pod", func() {
+			config := new(corev1.Secret)
 			Eventually(
 				func(g Gomega) {
 					err := k8sClient.Get(
 						ctx,
 						client.ObjectKey{
-							Name:      autoscalingListener.Name,
+							Name:      scaleSetListenerConfigName(autoscalingListener),
 							Namespace: autoscalingListener.Namespace,
 						},
-						pod,
+						config,
 					)
 
 					g.Expect(err).NotTo(HaveOccurred(), "failed to get pod")
-					g.Expect(pod.Spec.Containers).NotTo(BeEmpty(), "pod should have containers")
-					g.Expect(pod.Spec.Containers[0].Env).NotTo(BeEmpty(), "pod should have env variables")
 
-					var env *corev1.EnvVar
-					for _, e := range pod.Spec.Containers[0].Env {
-						if e.Name == "GITHUB_SERVER_ROOT_CA" {
-							env = &e
-							break
-						}
-					}
-					g.Expect(env).NotTo(BeNil(), "pod should have an env variable named GITHUB_SERVER_ROOT_CA_PATH")
+					g.Expect(config.Data["config.json"]).ToNot(BeEmpty(), "listener configuration file should not be empty")
+
+					var listenerConfig listenerconfig.Config
+					err = json.Unmarshal(config.Data["config.json"], &listenerConfig)
+					g.Expect(err).NotTo(HaveOccurred(), "failed to parse listener configuration file")
 
 					cert, err := os.ReadFile(filepath.Join(
 						"../../",
@@ -1000,7 +1056,7 @@ var _ = Describe("Test GitHub Server TLS configuration", func() {
 					))
 					g.Expect(err).NotTo(HaveOccurred(), "failed to read rootCA.crt")
 
-					g.Expect(env.Value).To(
+					g.Expect(listenerConfig.ServerRootCA).To(
 						BeEquivalentTo(string(cert)),
 						"GITHUB_SERVER_ROOT_CA should be the rootCA.crt",
 					)

controllers/actions.github.com/autoscalingrunnerset_controller.go

@@ -39,14 +39,17 @@ import (
 	"sigs.k8s.io/controller-runtime/pkg/handler"
 	"sigs.k8s.io/controller-runtime/pkg/predicate"
 	"sigs.k8s.io/controller-runtime/pkg/reconcile"
-	"sigs.k8s.io/controller-runtime/pkg/source"
 )
 
 const (
-	labelKeyRunnerSpecHash            = "runner-spec-hash"
+	annotationKeyRunnerSpecHash = "actions.github.com/runner-spec-hash"
+	// annotationKeyValuesHash is hash of the entire values json.
+	// This is used to determine if the values have changed, so we can
+	// re-create listener.
+	annotationKeyValuesHash = "actions.github.com/values-hash"
+
 	autoscalingRunnerSetFinalizerName = "autoscalingrunnerset.actions.github.com/finalizer"
 	runnerScaleSetIdAnnotationKey     = "runner-scale-set-id"
-	runnerScaleSetNameAnnotationKey   = "runner-scale-set-name"
 )
 
 type UpdateStrategy string
@@ -206,7 +209,7 @@ func (r *AutoscalingRunnerSetReconciler) Reconcile(ctx context.Context, req ctrl
 	}
 
 	// Make sure the runner scale set name is up to date
-	currentRunnerScaleSetName, ok := autoscalingRunnerSet.Annotations[runnerScaleSetNameAnnotationKey]
+	currentRunnerScaleSetName, ok := autoscalingRunnerSet.Annotations[AnnotationKeyGitHubRunnerScaleSetName]
 	if !ok || (len(autoscalingRunnerSet.Spec.RunnerScaleSetName) > 0 && !strings.EqualFold(currentRunnerScaleSetName, autoscalingRunnerSet.Spec.RunnerScaleSetName)) {
 		log.Info("AutoScalingRunnerSet runner scale set name changed. Updating the runner scale set.")
 		return r.updateRunnerScaleSetName(ctx, autoscalingRunnerSet, log)
@@ -232,9 +235,8 @@ func (r *AutoscalingRunnerSetReconciler) Reconcile(ctx context.Context, req ctrl
 		return r.createEphemeralRunnerSet(ctx, autoscalingRunnerSet, log)
 	}
 
-	desiredSpecHash := autoscalingRunnerSet.RunnerSetSpecHash()
 	for _, runnerSet := range existingRunnerSets.all() {
-		log.Info("Find existing ephemeral runner set", "name", runnerSet.Name, "specHash", runnerSet.Labels[labelKeyRunnerSpecHash])
+		log.Info("Find existing ephemeral runner set", "name", runnerSet.Name, "specHash", runnerSet.Annotations[annotationKeyRunnerSpecHash])
 	}
 
 	// Make sure the AutoscalingListener is up and running in the controller namespace
@@ -251,7 +253,9 @@ func (r *AutoscalingRunnerSetReconciler) Reconcile(ctx context.Context, req ctrl
 	}
 
 	// Our listener pod is out of date, so we need to delete it to get a new recreate.
-	if listenerFound && (listener.Labels[labelKeyRunnerSpecHash] != autoscalingRunnerSet.ListenerSpecHash()) {
+	listenerValuesHashChanged := listener.Annotations[annotationKeyValuesHash] != autoscalingRunnerSet.Annotations[annotationKeyValuesHash]
+	listenerSpecHashChanged := listener.Annotations[annotationKeyRunnerSpecHash] != autoscalingRunnerSet.ListenerSpecHash()
+	if listenerFound && (listenerValuesHashChanged || listenerSpecHashChanged) {
 		log.Info("RunnerScaleSetListener is out of date. Deleting it so that it is recreated", "name", listener.Name)
 		if err := r.Delete(ctx, listener); err != nil {
 			if kerrors.IsNotFound(err) {
@@ -265,7 +269,7 @@ func (r *AutoscalingRunnerSetReconciler) Reconcile(ctx context.Context, req ctrl
 		return ctrl.Result{}, nil
 	}
 
-	if desiredSpecHash != latestRunnerSet.Labels[labelKeyRunnerSpecHash] {
+	if latestRunnerSet.Annotations[annotationKeyRunnerSpecHash] != autoscalingRunnerSet.RunnerSetSpecHash() {
 		if r.drainingJobs(&latestRunnerSet.Status) {
 			log.Info("Latest runner set spec hash does not match the current autoscaling runner set. Waiting for the running and pending runners to finish:", "running", latestRunnerSet.Status.RunningEphemeralRunners, "pending", latestRunnerSet.Status.PendingEphemeralRunners)
 			log.Info("Scaling down the number of desired replicas to 0")
@@ -273,6 +277,7 @@ func (r *AutoscalingRunnerSetReconciler) Reconcile(ctx context.Context, req ctrl
 			// need to scale down to 0
 			err := patch(ctx, r.Client, latestRunnerSet, func(obj *v1alpha1.EphemeralRunnerSet) {
 				obj.Spec.Replicas = 0
+				obj.Spec.PatchID = 0
 			})
 			if err != nil {
 				log.Error(err, "Failed to patch runner set to set desired count to 0")
@@ -467,6 +472,8 @@ func (r *AutoscalingRunnerSetReconciler) createRunnerScaleSet(ctx context.Contex
 		Version:    build.Version,
 		CommitSHA:  build.CommitSHA,
 		ScaleSetID: runnerScaleSet.Id,
+		HasProxy:   autoscalingRunnerSet.Spec.Proxy != nil,
+		Subsystem:  "controller",
 	})
 
 	logger.Info("Created/Reused a runner scale set", "id", runnerScaleSet.Id, "runnerGroupName", runnerScaleSet.RunnerGroupName)
@@ -479,7 +486,7 @@ func (r *AutoscalingRunnerSetReconciler) createRunnerScaleSet(ctx context.Contex
 
 	logger.Info("Adding runner scale set ID, name and runner group name as an annotation and url labels")
 	if err = patch(ctx, r.Client, autoscalingRunnerSet, func(obj *v1alpha1.AutoscalingRunnerSet) {
-		obj.Annotations[runnerScaleSetNameAnnotationKey] = runnerScaleSet.Name
+		obj.Annotations[AnnotationKeyGitHubRunnerScaleSetName] = runnerScaleSet.Name
 		obj.Annotations[runnerScaleSetIdAnnotationKey] = strconv.Itoa(runnerScaleSet.Id)
 		obj.Annotations[AnnotationKeyGitHubRunnerGroupName] = runnerScaleSet.RunnerGroupName
 		if err := applyGitHubURLLabels(obj.Spec.GitHubConfigUrl, obj.Labels); err != nil { // should never happen
@@ -527,9 +534,10 @@ func (r *AutoscalingRunnerSetReconciler) updateRunnerScaleSetRunnerGroup(ctx con
 		return ctrl.Result{}, err
 	}
 
-	logger.Info("Updating runner scale set runner group name as an annotation")
+	logger.Info("Updating runner scale set name and runner group name as annotations")
 	if err := patch(ctx, r.Client, autoscalingRunnerSet, func(obj *v1alpha1.AutoscalingRunnerSet) {
 		obj.Annotations[AnnotationKeyGitHubRunnerGroupName] = updatedRunnerScaleSet.RunnerGroupName
+		obj.Annotations[AnnotationKeyGitHubRunnerScaleSetName] = updatedRunnerScaleSet.Name
 	}); err != nil {
 		logger.Error(err, "Failed to update runner group name annotation")
 		return ctrl.Result{}, err
@@ -565,7 +573,7 @@ func (r *AutoscalingRunnerSetReconciler) updateRunnerScaleSetName(ctx context.Co
 
 	logger.Info("Updating runner scale set name as an annotation")
 	if err := patch(ctx, r.Client, autoscalingRunnerSet, func(obj *v1alpha1.AutoscalingRunnerSet) {
-		obj.Annotations[runnerScaleSetNameAnnotationKey] = updatedRunnerScaleSet.Name
+		obj.Annotations[AnnotationKeyGitHubRunnerScaleSetName] = updatedRunnerScaleSet.Name
 	}); err != nil {
 		logger.Error(err, "Failed to update runner scale set name annotation")
 		return ctrl.Result{}, err
@@ -774,8 +782,8 @@ func (r *AutoscalingRunnerSetReconciler) SetupWithManager(mgr ctrl.Manager) erro
 	return ctrl.NewControllerManagedBy(mgr).
 		For(&v1alpha1.AutoscalingRunnerSet{}).
 		Owns(&v1alpha1.EphemeralRunnerSet{}).
-		Watches(&source.Kind{Type: &v1alpha1.AutoscalingListener{}}, handler.EnqueueRequestsFromMapFunc(
-			func(o client.Object) []reconcile.Request {
+		Watches(&v1alpha1.AutoscalingListener{}, handler.EnqueueRequestsFromMapFunc(
+			func(_ context.Context, o client.Object) []reconcile.Request {
 				autoscalingListener := o.(*v1alpha1.AutoscalingListener)
 				return []reconcile.Request{
 					{

controllers/actions.github.com/autoscalingrunnerset_controller_test.go

@@ -280,6 +280,10 @@ var _ = Describe("Test AutoScalingRunnerSet controller", Ordered, func() {
 			// This should trigger re-creation of EphemeralRunnerSet and Listener
 			patched := autoscalingRunnerSet.DeepCopy()
 			patched.Spec.Template.Spec.PriorityClassName = "test-priority-class"
+			if patched.ObjectMeta.Annotations == nil {
+				patched.ObjectMeta.Annotations = make(map[string]string)
+			}
+			patched.ObjectMeta.Annotations[annotationKeyValuesHash] = "test-hash"
 			err = k8sClient.Patch(ctx, patched, client.MergeFrom(autoscalingRunnerSet))
 			Expect(err).NotTo(HaveOccurred(), "failed to patch AutoScalingRunnerSet")
 			autoscalingRunnerSet = patched.DeepCopy()
@@ -297,10 +301,10 @@ var _ = Describe("Test AutoScalingRunnerSet controller", Ordered, func() {
 						return "", fmt.Errorf("We should have only 1 EphemeralRunnerSet, but got %v", len(runnerSetList.Items))
 					}
 
-					return runnerSetList.Items[0].Labels[labelKeyRunnerSpecHash], nil
+					return runnerSetList.Items[0].Annotations[annotationKeyRunnerSpecHash], nil
 				},
 				autoscalingRunnerSetTestTimeout,
-				autoscalingRunnerSetTestInterval).ShouldNot(BeEquivalentTo(runnerSet.Labels[labelKeyRunnerSpecHash]), "New EphemeralRunnerSet should be created")
+				autoscalingRunnerSetTestInterval).ShouldNot(BeEquivalentTo(runnerSet.Annotations[annotationKeyRunnerSpecHash]), "New EphemeralRunnerSet should be created")
 
 			// We should create a new listener
 			Eventually(
@@ -334,6 +338,55 @@ var _ = Describe("Test AutoScalingRunnerSet controller", Ordered, func() {
 			err = k8sClient.Patch(ctx, patched, client.MergeFrom(autoscalingRunnerSet))
 			Expect(err).NotTo(HaveOccurred(), "failed to patch AutoScalingRunnerSet")
 
+			// We should not re-create a new EphemeralRunnerSet
+			Consistently(
+				func() (string, error) {
+					runnerSetList := new(v1alpha1.EphemeralRunnerSetList)
+					err := k8sClient.List(ctx, runnerSetList, client.InNamespace(autoscalingRunnerSet.Namespace))
+					if err != nil {
+						return "", err
+					}
+
+					if len(runnerSetList.Items) != 1 {
+						return "", fmt.Errorf("We should have only 1 EphemeralRunnerSet, but got %v", len(runnerSetList.Items))
+					}
+
+					return string(runnerSetList.Items[0].UID), nil
+				},
+				autoscalingRunnerSetTestTimeout,
+				autoscalingRunnerSetTestInterval).Should(BeEquivalentTo(string(runnerSet.UID)), "New EphemeralRunnerSet should not be created")
+
+			// We should only re-create a new listener
+			Eventually(
+				func() (string, error) {
+					listener := new(v1alpha1.AutoscalingListener)
+					err := k8sClient.Get(ctx, client.ObjectKey{Name: scaleSetListenerName(autoscalingRunnerSet), Namespace: autoscalingRunnerSet.Namespace}, listener)
+					if err != nil {
+						return "", err
+					}
+
+					return string(listener.UID), nil
+				},
+				autoscalingRunnerSetTestTimeout,
+				autoscalingRunnerSetTestInterval).ShouldNot(BeEquivalentTo(string(listener.UID)), "New Listener should be created")
+
+			// Only update the values hash for the autoscaling runner set
+			// This should trigger re-creation of the Listener only
+			runnerSetList = new(v1alpha1.EphemeralRunnerSetList)
+			err = k8sClient.List(ctx, runnerSetList, client.InNamespace(autoscalingRunnerSet.Namespace))
+			Expect(err).NotTo(HaveOccurred(), "failed to list EphemeralRunnerSet")
+			Expect(len(runnerSetList.Items)).To(Equal(1), "There should be 1 EphemeralRunnerSet")
+			runnerSet = runnerSetList.Items[0]
+
+			listener = new(v1alpha1.AutoscalingListener)
+			err = k8sClient.Get(ctx, client.ObjectKey{Name: scaleSetListenerName(autoscalingRunnerSet), Namespace: autoscalingRunnerSet.Namespace}, listener)
+			Expect(err).NotTo(HaveOccurred(), "failed to get Listener")
+
+			patched = autoscalingRunnerSet.DeepCopy()
+			patched.ObjectMeta.Annotations[annotationKeyValuesHash] = "hash-changes"
+			err = k8sClient.Patch(ctx, patched, client.MergeFrom(autoscalingRunnerSet))
+			Expect(err).NotTo(HaveOccurred(), "failed to patch AutoScalingRunnerSet")
+
 			// We should not re-create a new EphemeralRunnerSet
 			Consistently(
 				func() (string, error) {
@@ -493,6 +546,10 @@ var _ = Describe("Test AutoScalingRunnerSet controller", Ordered, func() {
 			// Patch the AutoScalingRunnerSet image which should trigger
 			// the recreation of the Listener and EphemeralRunnerSet
 			patched := autoscalingRunnerSet.DeepCopy()
+			if patched.ObjectMeta.Annotations == nil {
+				patched.ObjectMeta.Annotations = make(map[string]string)
+			}
+			patched.ObjectMeta.Annotations[annotationKeyValuesHash] = "testgroup2"
 			patched.Spec.Template.Spec = corev1.PodSpec{
 				Containers: []corev1.Container{
 					{
@@ -501,7 +558,6 @@ var _ = Describe("Test AutoScalingRunnerSet controller", Ordered, func() {
 					},
 				},
 			}
-			// patched.Spec.Template.Spec.PriorityClassName = "test-priority-class"
 			err = k8sClient.Patch(ctx, patched, client.MergeFrom(autoscalingRunnerSet))
 			Expect(err).NotTo(HaveOccurred(), "failed to patch AutoScalingRunnerSet")
 			autoscalingRunnerSet = patched.DeepCopy()
@@ -698,7 +754,7 @@ var _ = Describe("Test AutoScalingController updates", Ordered, func() {
 						return "", err
 					}
 
-					if val, ok := ars.Annotations[runnerScaleSetNameAnnotationKey]; ok {
+					if val, ok := ars.Annotations[AnnotationKeyGitHubRunnerScaleSetName]; ok {
 						return val, nil
 					}
 
@@ -722,7 +778,7 @@ var _ = Describe("Test AutoScalingController updates", Ordered, func() {
 						return "", err
 					}
 
-					if val, ok := ars.Annotations[runnerScaleSetNameAnnotationKey]; ok {
+					if val, ok := ars.Annotations[AnnotationKeyGitHubRunnerScaleSetName]; ok {
 						return val, nil
 					}
 

controllers/actions.github.com/constants.go

@@ -2,7 +2,6 @@ package actionsgithubcom
 
 import (
 	"github.com/actions/actions-runner-controller/logging"
-	corev1 "k8s.io/api/core/v1"
 )
 
 const (
@@ -40,7 +39,11 @@ const (
 // Finalizer used to protect resources from deletion while AutoscalingRunnerSet is running
 const AutoscalingRunnerSetCleanupFinalizerName = "actions.github.com/cleanup-protection"
 
-const AnnotationKeyGitHubRunnerGroupName = "actions.github.com/runner-group-name"
+const (
+	AnnotationKeyGitHubRunnerGroupName    = "actions.github.com/runner-group-name"
+	AnnotationKeyGitHubRunnerScaleSetName = "actions.github.com/runner-scale-set-name"
+	AnnotationKeyPatchID                  = "actions.github.com/patch-id"
+)
 
 // Labels applied to listener roles
 const (
@@ -59,10 +62,6 @@ const (
 	AnnotationKeyNoPermissionServiceAccountName   = "actions.github.com/cleanup-no-permission-service-account-name"
 )
 
-// DefaultScaleSetListenerImagePullPolicy is the default pull policy applied
-// to the listener when ImagePullPolicy is not specified
-const DefaultScaleSetListenerImagePullPolicy = corev1.PullIfNotPresent
-
 // DefaultScaleSetListenerLogLevel is the default log level applied
 const DefaultScaleSetListenerLogLevel = string(logging.LogLevelDebug)
 
@@ -71,3 +70,9 @@ const DefaultScaleSetListenerLogFormat = string(logging.LogFormatText)
 
 // ownerKey is field selector matching the owner name of a particular resource
 const resourceOwnerKey = ".metadata.controller"
+
+// EphemeralRunner pod creation failure reasons
+const (
+	ReasonTooManyPodFailures = "TooManyPodFailures"
+	ReasonInvalidPodFailure  = "InvalidPod"
+)

controllers/actions.github.com/ephemeralrunner_controller.go

@@ -21,7 +21,6 @@ import (
 	"errors"
 	"fmt"
 	"net/http"
-	"strings"
 	"time"
 
 	"github.com/actions/actions-runner-controller/apis/actions.github.com/v1alpha1"
@@ -133,6 +132,23 @@ func (r *EphemeralRunnerReconciler) Reconcile(ctx context.Context, req ctrl.Requ
 		return ctrl.Result{}, nil
 	}
 
+	if ephemeralRunner.IsDone() {
+		log.Info("Cleaning up resources after after ephemeral runner termination", "phase", ephemeralRunner.Status.Phase)
+		done, err := r.cleanupResources(ctx, ephemeralRunner, log)
+		if err != nil {
+			log.Error(err, "Failed to clean up ephemeral runner owned resources")
+			return ctrl.Result{}, err
+		}
+		if !done {
+			log.Info("Waiting for ephemeral runner owned resources to be deleted")
+			return ctrl.Result{Requeue: true}, nil
+		}
+		// Stop reconciling on this object.
+		// The EphemeralRunnerSet is responsible for cleaning it up.
+		log.Info("EphemeralRunner has already finished. Stopping reconciliation and waiting for EphemeralRunnerSet to clean it up", "phase", ephemeralRunner.Status.Phase)
+		return ctrl.Result{}, nil
+	}
+
 	if !controllerutil.ContainsFinalizer(ephemeralRunner, ephemeralRunnerActionsFinalizerName) {
 		log.Info("Adding runner registration finalizer")
 		err := patch(ctx, r.Client, ephemeralRunner, func(obj *v1alpha1.EphemeralRunner) {
@@ -159,13 +175,6 @@ func (r *EphemeralRunnerReconciler) Reconcile(ctx context.Context, req ctrl.Requ
 		return ctrl.Result{}, nil
 	}
 
-	if ephemeralRunner.Status.Phase == corev1.PodSucceeded || ephemeralRunner.Status.Phase == corev1.PodFailed {
-		// Stop reconciling on this object.
-		// The EphemeralRunnerSet is responsible for cleaning it up.
-		log.Info("EphemeralRunner has already finished. Stopping reconciliation and waiting for EphemeralRunnerSet to clean it up", "phase", ephemeralRunner.Status.Phase)
-		return ctrl.Result{}, nil
-	}
-
 	if ephemeralRunner.Status.RunnerId == 0 {
 		log.Info("Creating new ephemeral runner registration and updating status with runner config")
 		return r.updateStatusWithRunnerConfig(ctx, ephemeralRunner, log)
@@ -191,7 +200,8 @@ func (r *EphemeralRunnerReconciler) Reconcile(ctx context.Context, req ctrl.Requ
 
 		case len(ephemeralRunner.Status.Failures) > 5:
 			log.Info("EphemeralRunner has failed more than 5 times. Marking it as failed")
-			if err := r.markAsFailed(ctx, ephemeralRunner, log); err != nil {
+			errMessage := fmt.Sprintf("Pod has failed to start more than 5 times: %s", pod.Status.Message)
+			if err := r.markAsFailed(ctx, ephemeralRunner, errMessage, ReasonTooManyPodFailures, log); err != nil {
 				log.Error(err, "Failed to set ephemeral runner to phase Failed")
 				return ctrl.Result{}, err
 			}
@@ -200,7 +210,22 @@ func (r *EphemeralRunnerReconciler) Reconcile(ctx context.Context, req ctrl.Requ
 		default:
 			// Pod was not found. Create if the pod has never been created
 			log.Info("Creating new EphemeralRunner pod.")
-			return r.createPod(ctx, ephemeralRunner, secret, log)
+			result, err := r.createPod(ctx, ephemeralRunner, secret, log)
+			switch {
+			case err == nil:
+				return result, nil
+			case kerrors.IsInvalid(err) || kerrors.IsForbidden(err):
+				log.Error(err, "Failed to create a pod due to unrecoverable failure")
+				errMessage := fmt.Sprintf("Failed to create the pod: %v", err)
+				if err := r.markAsFailed(ctx, ephemeralRunner, errMessage, ReasonInvalidPodFailure, log); err != nil {
+					log.Error(err, "Failed to set ephemeral runner to phase Failed")
+					return ctrl.Result{}, err
+				}
+				return ctrl.Result{}, nil
+			default:
+				log.Error(err, "Failed to create the pod")
+				return ctrl.Result{}, err
+			}
 		}
 	}
 
@@ -269,14 +294,17 @@ func (r *EphemeralRunnerReconciler) Reconcile(ctx context.Context, req ctrl.Requ
 }
 
 func (r *EphemeralRunnerReconciler) cleanupRunnerFromService(ctx context.Context, ephemeralRunner *v1alpha1.EphemeralRunner, log logr.Logger) (ctrl.Result, error) {
+	if err := r.deleteRunnerFromService(ctx, ephemeralRunner, log); err != nil {
 		actionsError := &actions.ActionsError{}
-	err := r.deleteRunnerFromService(ctx, ephemeralRunner, log)
-	if err != nil {
-		if errors.As(err, &actionsError) &&
-			actionsError.StatusCode == http.StatusBadRequest &&
-			strings.Contains(actionsError.ExceptionName, "JobStillRunningException") {
+		if !errors.As(err, &actionsError) {
+			log.Error(err, "Failed to clean up runner from the service (not an ActionsError)")
+			return ctrl.Result{}, err
+		}
+
+		if actionsError.StatusCode == http.StatusBadRequest && actionsError.IsException("JobStillRunningException") {
 			log.Info("Runner is still running the job. Re-queue in 30 seconds")
 			return ctrl.Result{RequeueAfter: 30 * time.Second}, nil
+
 		}
 
 		log.Error(err, "Failed clean up runner from the service")
@@ -284,10 +312,9 @@ func (r *EphemeralRunnerReconciler) cleanupRunnerFromService(ctx context.Context
 	}
 
 	log.Info("Successfully removed runner registration from service")
-	err = patch(ctx, r.Client, ephemeralRunner, func(obj *v1alpha1.EphemeralRunner) {
+	if err := patch(ctx, r.Client, ephemeralRunner, func(obj *v1alpha1.EphemeralRunner) {
 		controllerutil.RemoveFinalizer(obj, ephemeralRunnerActionsFinalizerName)
-	})
-	if err != nil {
+	}); err != nil {
 		return ctrl.Result{}, err
 	}
 
@@ -308,7 +335,7 @@ func (r *EphemeralRunnerReconciler) cleanupResources(ctx context.Context, epheme
 			}
 		}
 		return false, nil
-	case err != nil && !kerrors.IsNotFound(err):
+	case !kerrors.IsNotFound(err):
 		return false, err
 	}
 	log.Info("Pod is deleted")
@@ -325,7 +352,7 @@ func (r *EphemeralRunnerReconciler) cleanupResources(ctx context.Context, epheme
 			}
 		}
 		return false, nil
-	case err != nil && !kerrors.IsNotFound(err):
+	case !kerrors.IsNotFound(err):
 		return false, err
 	}
 	log.Info("Secret is deleted")
@@ -423,12 +450,12 @@ func (r *EphemeralRunnerReconciler) cleanupRunnerLinkedSecrets(ctx context.Conte
 	return false, multierr.Combine(errs...)
 }
 
-func (r *EphemeralRunnerReconciler) markAsFailed(ctx context.Context, ephemeralRunner *v1alpha1.EphemeralRunner, log logr.Logger) error {
+func (r *EphemeralRunnerReconciler) markAsFailed(ctx context.Context, ephemeralRunner *v1alpha1.EphemeralRunner, errMessage string, reason string, log logr.Logger) error {
 	log.Info("Updating ephemeral runner status to Failed")
 	if err := patchSubResource(ctx, r.Status(), ephemeralRunner, func(obj *v1alpha1.EphemeralRunner) {
 		obj.Status.Phase = corev1.PodFailed
-		obj.Status.Reason = "TooManyPodFailures"
-		obj.Status.Message = "Pod has failed to start more than 5 times"
+		obj.Status.Reason = reason
+		obj.Status.Message = errMessage
 	}); err != nil {
 		return fmt.Errorf("failed to update ephemeral runner status Phase/Message: %v", err)
 	}
@@ -502,7 +529,7 @@ func (r *EphemeralRunnerReconciler) updateStatusWithRunnerConfig(ctx context.Con
 		}
 
 		if actionsError.StatusCode != http.StatusConflict ||
-			!strings.Contains(actionsError.ExceptionName, "AgentExistsException") {
+			!actionsError.IsException("AgentExistsException") {
 			return ctrl.Result{}, fmt.Errorf("failed to generate JIT config with Actions service error: %v", err)
 		}
 
@@ -659,7 +686,7 @@ func (r *EphemeralRunnerReconciler) updateRunStatusFromPod(ctx context.Context,
 		return nil
 	}
 
-	log.Info("Updating ephemeral runner status with pod phase", "phase", pod.Status.Phase, "reason", pod.Status.Reason, "message", pod.Status.Message)
+	log.Info("Updating ephemeral runner status with pod phase", "statusPhase", pod.Status.Phase, "statusReason", pod.Status.Reason, "statusMessage", pod.Status.Message)
 	err := patchSubResource(ctx, r.Status(), ephemeralRunner, func(obj *v1alpha1.EphemeralRunner) {
 		obj.Status.Phase = pod.Status.Phase
 		obj.Status.Ready = obj.Status.Ready || (pod.Status.Phase == corev1.PodRunning)
@@ -758,7 +785,7 @@ func (r EphemeralRunnerReconciler) runnerRegisteredWithService(ctx context.Conte
 		}
 
 		if actionsError.StatusCode != http.StatusNotFound ||
-			!strings.Contains(actionsError.ExceptionName, "AgentNotFoundException") {
+			!actionsError.IsException("AgentNotFoundException") {
 			return false, fmt.Errorf("failed to check if runner exists in GitHub service: %v", err)
 		}