Fix parsing AcquireJob MessageQueueTokenExpiredError (#2837)

.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: "0.4.0"
+  IMAGE_VERSION: "0.5.0"
 
 concurrency:
   # This will make sure we only apply the concurrency limits on pull requests

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

@@ -151,7 +151,7 @@ jobs:
           GHA_RUNNER_SCALE_SET_CONTROLLER_CHART_VERSION_TAG=$(cat charts/gha-runner-scale-set-controller/Chart.yaml | grep version: | cut -d " " -f 2)
           echo "GHA_RUNNER_SCALE_SET_CONTROLLER_CHART_VERSION_TAG=${GHA_RUNNER_SCALE_SET_CONTROLLER_CHART_VERSION_TAG}" >> $GITHUB_ENV
           helm package charts/gha-runner-scale-set-controller/ --version="${GHA_RUNNER_SCALE_SET_CONTROLLER_CHART_VERSION_TAG}"
-          helm push gha-rs-controller-"${GHA_RUNNER_SCALE_SET_CONTROLLER_CHART_VERSION_TAG}".tgz oci://ghcr.io/${{ steps.resolve_parameters.outputs.repository_owner }}/actions-runner-controller-charts
+          helm push gha-runner-scale-set-controller-"${GHA_RUNNER_SCALE_SET_CONTROLLER_CHART_VERSION_TAG}".tgz oci://ghcr.io/${{ steps.resolve_parameters.outputs.repository_owner }}/actions-runner-controller-charts
 
       - name: Job summary
         run: |
@@ -200,7 +200,7 @@ jobs:
           GHA_RUNNER_SCALE_SET_CHART_VERSION_TAG=$(cat charts/gha-runner-scale-set/Chart.yaml | grep version: | cut -d " " -f 2)
           echo "GHA_RUNNER_SCALE_SET_CHART_VERSION_TAG=${GHA_RUNNER_SCALE_SET_CHART_VERSION_TAG}" >> $GITHUB_ENV
           helm package charts/gha-runner-scale-set/ --version="${GHA_RUNNER_SCALE_SET_CHART_VERSION_TAG}"
-          helm push gha-rs-"${GHA_RUNNER_SCALE_SET_CHART_VERSION_TAG}".tgz oci://ghcr.io/${{ steps.resolve_parameters.outputs.repository_owner }}/actions-runner-controller-charts
+          helm push gha-runner-scale-set-"${GHA_RUNNER_SCALE_SET_CHART_VERSION_TAG}".tgz oci://ghcr.io/${{ steps.resolve_parameters.outputs.repository_owner }}/actions-runner-controller-charts
 
       - name: Job summary
         run: |

docs/gha-runner-scale-set-controller/samples/grafana-dashboard/README.md

@@ -0,0 +1,15 @@
+# Visualizing Autoscaling Runner Scale Set metrics with Grafana
+
+With metrics introduced in [gha-runner-scale-set-0.5.0](https://github.com/actions/actions-runner-controller/releases/tag/gha-runner-scale-set-0.5.0), you can now visualize the autoscaling behavior of your runner scale set with your tool of choice. This sample shows how to visualize the metrics with [Grafana](https://grafana.com/).
+
+## Demo
+
+![Grafana dashboard example](grafana-sample.png)
+
+## Setup
+
+We do not intend to provide a supported ARC dashboard. This is simply a reference and a demonstration for how you could leverage the metrics emitted by the controller-manager and listeners to visualize the autoscaling behavior of your runner scale set. We offer no promises of future upgrades to this sample.
+
+1. Make sure to have [Grafana](https://grafana.com/docs/grafana/latest/installation/) and [Prometheus](https://prometheus.io/docs/prometheus/latest/installation/) running in your cluster.
+2. Make sure that Prometheus is properly scraping the metrics endpoints of the controller-manager and listeners.
+3. Import the [dashboard](ARC-Autoscaling-Runner-Set-Monitoring_1692627561838.json.json) into Grafana.

github/actions/client.go

@@ -634,7 +634,18 @@ func (c *Client) AcquireJobs(ctx context.Context, runnerScaleSetId int, messageQ
 	}
 
 	if resp.StatusCode != http.StatusOK {
-		return nil, ParseActionsErrorFromResponse(resp)
+		if resp.StatusCode != http.StatusUnauthorized {
+			return nil, ParseActionsErrorFromResponse(resp)
+		}
+
+		defer resp.Body.Close()
+		body, err := io.ReadAll(resp.Body)
+		body = trimByteOrderMark(body)
+		if err != nil {
+			return nil, err
+		}
+
+		return nil, &MessageQueueTokenExpiredError{msg: string(body)}
 	}
 
 	var acquiredJobs *Int64List

github/actions/client_job_acquisition_test.go

@@ -2,6 +2,7 @@ package actions_test
 
 import (
 	"context"
+	"errors"
 	"net/http"
 	"strings"
 	"testing"
@@ -84,6 +85,39 @@ func TestAcquireJobs(t *testing.T) {
 		assert.NotNil(t, err)
 		assert.Equalf(t, actualRetry, expectedRetry, "A retry was expected after the first request but got: %v", actualRetry)
 	})
+
+	t.Run("Should return MessageQueueTokenExpiredError when http error is not Unauthorized", func(t *testing.T) {
+		want := []int64{1}
+
+		session := &actions.RunnerScaleSetSession{
+			RunnerScaleSet:          &actions.RunnerScaleSet{Id: 1},
+			MessageQueueAccessToken: "abc",
+		}
+		requestIDs := want
+
+		server := newActionsServer(t, http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+			if strings.HasSuffix(r.URL.Path, "/acquirablejobs") {
+				w.Write([]byte(`{"count": 1}`))
+				return
+			}
+			if r.Method == http.MethodPost {
+				http.Error(w, "Session expired", http.StatusUnauthorized)
+				return
+			}
+		}))
+
+		client, err := actions.NewClient(server.configURLForOrg("my-org"), auth)
+		require.NoError(t, err)
+
+		_, err = client.GetAcquirableJobs(ctx, 1)
+		require.NoError(t, err)
+
+		got, err := client.AcquireJobs(ctx, session.RunnerScaleSet.Id, session.MessageQueueAccessToken, requestIDs)
+		require.Error(t, err)
+		assert.Nil(t, got)
+		var expectedErr *actions.MessageQueueTokenExpiredError
+		assert.True(t, errors.As(err, &expectedErr))
+	})
 }
 
 func TestGetAcquirableJobs(t *testing.T) {