Re-use the last desired patch on empty batch (#3453)

cmd/ghalistener/worker/worker.go

@@ -41,7 +41,7 @@ type Worker struct {
 	clientset *kubernetes.Clientset
 	config    Config
 	lastPatch int
-	lastPatchID int
+	patchSeq  int
 	logger    *logr.Logger
 }
 
@@ -51,7 +51,7 @@ func New(config Config, options ...Option) (*Worker, error) {
 	w := &Worker{
 		config:    config,
 		lastPatch: -1,
-		lastPatchID: -1,
+		patchSeq:  -1,
 	}
 
 	conf, err := rest.InClusterConfig()
@@ -163,27 +163,8 @@ func (w *Worker) HandleJobStarted(ctx context.Context, jobInfo *actions.JobStart
 // 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) {
+func (w *Worker) HandleDesiredRunnerCount(ctx context.Context, count, jobsCompleted int) (int, error) {
-	// Max runners should always be set by the resource builder either to the configured value,
+	patchID := w.setDesiredWorkerState(count, jobsCompleted)
-	// 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{
@@ -200,8 +181,8 @@ func (w *Worker) HandleDesiredRunnerCount(ctx context.Context, count int, jobsCo
 	patch, err := json.Marshal(
 		&v1alpha1.EphemeralRunnerSet{
 			Spec: v1alpha1.EphemeralRunnerSetSpec{
-				Replicas: targetRunnerCount,
+				Replicas: w.lastPatch,
-				PatchID:  w.lastPatchID,
+				PatchID:  patchID,
 			},
 		},
 	)
@@ -210,14 +191,13 @@ func (w *Worker) HandleDesiredRunnerCount(ctx context.Context, count int, jobsCo
 		return 0, err
 	}
 
+	w.logger.Info("Compare", "original", string(original), "patch", string(patch))
 	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("Preparing EphemeralRunnerSet update", "json", string(mergePatch))
-
-	w.logger.Info("Scaling ephemeral runner set", logValues...)
 
 	patchedEphemeralRunnerSet := &v1alpha1.EphemeralRunnerSet{}
 	err = w.clientset.RESTClient().
@@ -238,5 +218,40 @@ func (w *Worker) HandleDesiredRunnerCount(ctx context.Context, count int, jobsCo
 		"name", w.config.EphemeralRunnerSetName,
 		"replicas", patchedEphemeralRunnerSet.Spec.Replicas,
 	)
-	return targetRunnerCount, nil
+	return w.lastPatch, nil
+}
+
+// calculateDesiredState calculates the desired state of the worker based on the desired count and the the number of jobs completed.
+func (w *Worker) setDesiredWorkerState(count, jobsCompleted int) int {
+	// 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)
+	w.patchSeq++
+	desiredPatchID := w.patchSeq
+
+	if count == 0 && jobsCompleted == 0 { // empty batch
+		targetRunnerCount = max(w.lastPatch, targetRunnerCount)
+		if targetRunnerCount == w.config.MinRunners {
+			// We have an empty batch, and the last patch was the min runners.
+			// Since this is an empty batch, and we are at the min runners, they should all be idle.
+			// If controller created few more pods on accident (during scale down events),
+			// this situation allows the controller to scale down to the min runners.
+			// However, it is important to keep the patch sequence increasing so we don't ignore one batch.
+			desiredPatchID = 0
+		}
+	}
+
+	w.lastPatch = targetRunnerCount
+
+	w.logger.Info(
+		"Calculated target runner count",
+		"assigned job", count,
+		"decision", targetRunnerCount,
+		"min", w.config.MinRunners,
+		"max", w.config.MaxRunners,
+		"currentRunnerCount", w.lastPatch,
+		"jobsCompleted", jobsCompleted,
+	)
+
+	return desiredPatchID
 }

cmd/ghalistener/worker/worker_test.go

@@ -0,0 +1,326 @@
+package worker
+
+import (
+	"math"
+	"testing"
+
+	"github.com/go-logr/logr"
+	"github.com/stretchr/testify/assert"
+)
+
+func TestSetDesiredWorkerState_MinMaxDefaults(t *testing.T) {
+	logger := logr.Discard()
+	newEmptyWorker := func() *Worker {
+		return &Worker{
+			config: Config{
+				MinRunners: 0,
+				MaxRunners: math.MaxInt32,
+			},
+			lastPatch: -1,
+			patchSeq:  -1,
+			logger:    &logger,
+		}
+	}
+
+	t.Run("init calculate with acquired 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+		assert.Equal(t, 0, patchID)
+	})
+
+	t.Run("init calculate with acquired 1", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(1, 0)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+		assert.Equal(t, 0, patchID)
+	})
+
+	t.Run("increment patch when job done", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(1, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 1)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("increment patch when called with same parameters", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(1, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(1, 0)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("calculate desired scale when acquired > 0 and completed > 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(1, 1)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("re-use the last state when acquired == 0 and completed == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(1, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("adjust when acquired == 0 and completed == 1", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(1, 1)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 1)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+}
+
+func TestSetDesiredWorkerState_MinSet(t *testing.T) {
+	logger := logr.Discard()
+	newEmptyWorker := func() *Worker {
+		return &Worker{
+			config: Config{
+				MinRunners: 1,
+				MaxRunners: math.MaxInt32,
+			},
+			lastPatch: -1,
+			patchSeq:  -1,
+			logger:    &logger,
+		}
+	}
+
+	t.Run("initial scale when acquired == 0 and completed == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("re-use the old state on count == 0 and completed == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(2, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 3, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("request back to 0 on job done", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(2, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 1)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("desired patch is 0 but sequence continues on empty batch and min runners", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(3, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 4, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+
+		patchID = w.setDesiredWorkerState(0, 3)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+
+		// Empty batch on min runners
+		patchID = w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 0, patchID) // forcing the state
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 2, w.patchSeq)
+	})
+
+}
+
+func TestSetDesiredWorkerState_MaxSet(t *testing.T) {
+	logger := logr.Discard()
+	newEmptyWorker := func() *Worker {
+		return &Worker{
+			config: Config{
+				MinRunners: 0,
+				MaxRunners: 5,
+			},
+			lastPatch: -1,
+			patchSeq:  -1,
+			logger:    &logger,
+		}
+	}
+
+	t.Run("initial scale when acquired == 0 and completed == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("re-use the old state on count == 0 and completed == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(2, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 2, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("request back to 0 on job done", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(2, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 1)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("scale up to max when count > max", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(6, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 5, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("scale to max when count == max", func(t *testing.T) {
+		w := newEmptyWorker()
+		w.setDesiredWorkerState(5, 0)
+		assert.Equal(t, 5, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("scale to max when count > max and completed > 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(1, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(6, 1)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 5, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("scale back to 0 when count was > max", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(6, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 1)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("force 0 on empty batch and last patch == min runners", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(3, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 3, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+
+		patchID = w.setDesiredWorkerState(0, 3)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+
+		// Empty batch on min runners
+		patchID = w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 0, patchID) // forcing the state
+		assert.Equal(t, 0, w.lastPatch)
+		assert.Equal(t, 2, w.patchSeq)
+	})
+}
+
+func TestSetDesiredWorkerState_MinMaxSet(t *testing.T) {
+	logger := logr.Discard()
+	newEmptyWorker := func() *Worker {
+		return &Worker{
+			config: Config{
+				MinRunners: 1,
+				MaxRunners: 3,
+			},
+			lastPatch: -1,
+			patchSeq:  -1,
+			logger:    &logger,
+		}
+	}
+
+	t.Run("initial scale when acquired == 0 and completed == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("re-use the old state on count == 0 and completed == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(2, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 3, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("scale to min when count == 0", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(2, 0)
+		assert.Equal(t, 0, patchID)
+		patchID = w.setDesiredWorkerState(0, 1)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+	})
+
+	t.Run("scale up to max when count > max", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(4, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 3, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("scale to max when count == max", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(3, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 3, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+	})
+
+	t.Run("force 0 on empty batch and last patch == min runners", func(t *testing.T) {
+		w := newEmptyWorker()
+		patchID := w.setDesiredWorkerState(3, 0)
+		assert.Equal(t, 0, patchID)
+		assert.Equal(t, 3, w.lastPatch)
+		assert.Equal(t, 0, w.patchSeq)
+
+		patchID = w.setDesiredWorkerState(0, 3)
+		assert.Equal(t, 1, patchID)
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 1, w.patchSeq)
+
+		// Empty batch on min runners
+		patchID = w.setDesiredWorkerState(0, 0)
+		assert.Equal(t, 0, patchID) // forcing the state
+		assert.Equal(t, 1, w.lastPatch)
+		assert.Equal(t, 2, w.patchSeq)
+	})
+}

controllers/actions.github.com/ephemeralrunnerset_controller.go

@@ -213,9 +213,6 @@ func (r *EphemeralRunnerSetReconciler) Reconcile(ctx context.Context, req ctrl.R
 			// on the next batch
 		case ephemeralRunnerSet.Spec.PatchID == 0 && total > ephemeralRunnerSet.Spec.Replicas:
 			count := total - ephemeralRunnerSet.Spec.Replicas
-			if count <= 0 {
-				break
-			}
 			log.Info("Deleting ephemeral runners (scale down)", "count", count)
 			if err := r.deleteIdleEphemeralRunners(
 				ctx,