Merge pull request #112003 from sanposhiho/metrics-goroutine

feature(scheduler): add "goroutines" metric and deprecate the "scheduler_goroutines" metric

pkg/scheduler/framework/parallelize/parallelism.go

@@ -21,6 +21,7 @@ import (
 	"math"
 
 	"k8s.io/client-go/util/workqueue"
+	"k8s.io/kubernetes/pkg/scheduler/metrics"
 )
 
 // DefaultParallelism is the default parallelism used in scheduler.
@@ -51,6 +52,13 @@ func chunkSizeFor(n, parallelism int) int {
 }
 
 // Until is a wrapper around workqueue.ParallelizeUntil to use in scheduling algorithms.
-func (p Parallelizer) Until(ctx context.Context, pieces int, doWorkPiece workqueue.DoWorkPieceFunc) {
+// A given operation will be a label that is recorded in the goroutine metric.
-	workqueue.ParallelizeUntil(ctx, p.parallelism, pieces, doWorkPiece, workqueue.WithChunkSize(chunkSizeFor(pieces, p.parallelism)))
+func (p Parallelizer) Until(ctx context.Context, pieces int, doWorkPiece workqueue.DoWorkPieceFunc, operation string) {
+	withMetrics := func(piece int) {
+		metrics.Goroutines.WithLabelValues(operation).Inc()
+		defer metrics.Goroutines.WithLabelValues(operation).Dec()
+		doWorkPiece(piece)
+	}
+
+	workqueue.ParallelizeUntil(ctx, p.parallelism, pieces, withMetrics, workqueue.WithChunkSize(chunkSizeFor(pieces, p.parallelism)))
 }

pkg/scheduler/framework/plugins/interpodaffinity/filtering.go

@@ -170,7 +170,7 @@ func (pl *InterPodAffinity) getExistingAntiAffinityCounts(ctx context.Context, p
 			topoMaps[atomic.AddInt32(&index, 1)] = topoMap
 		}
 	}
-	pl.parallelizer.Until(ctx, len(nodes), processNode)
+	pl.parallelizer.Until(ctx, len(nodes), processNode, pl.Name())
 
 	result := make(topologyToMatchedTermCount)
 	for i := 0; i <= int(index); i++ {
@@ -216,7 +216,7 @@ func (pl *InterPodAffinity) getIncomingAffinityAntiAffinityCounts(ctx context.Co
 			antiAffinityCountsList[k] = antiAffinity
 		}
 	}
-	pl.parallelizer.Until(ctx, len(allNodes), processNode)
+	pl.parallelizer.Until(ctx, len(allNodes), processNode, pl.Name())
 
 	for i := 0; i <= int(index); i++ {
 		affinityCounts.append(affinityCountsList[i])

pkg/scheduler/framework/plugins/interpodaffinity/scoring.go

@@ -203,7 +203,7 @@ func (pl *InterPodAffinity) PreScore(
 			topoScores[atomic.AddInt32(&index, 1)] = topoScore
 		}
 	}
-	pl.parallelizer.Until(pCtx, len(allNodes), processNode)
+	pl.parallelizer.Until(pCtx, len(allNodes), processNode, pl.Name())
 
 	for i := 0; i <= int(index); i++ {
 		state.topologyScore.append(topoScores[i])

pkg/scheduler/framework/plugins/podtopologyspread/filtering.go

@@ -304,7 +304,7 @@ func (pl *PodTopologySpread) calPreFilterState(ctx context.Context, pod *v1.Pod)
 		}
 		tpCountsByNode[i] = tpCounts
 	}
-	pl.parallelizer.Until(ctx, len(allNodes), processNode)
+	pl.parallelizer.Until(ctx, len(allNodes), processNode, pl.Name())
 
 	for _, tpCounts := range tpCountsByNode {
 		for tp, count := range tpCounts {

pkg/scheduler/framework/plugins/podtopologyspread/filtering_test.go

@@ -2013,7 +2013,7 @@ func BenchmarkFilter(b *testing.B) {
 					n, _ := p.sharedLister.NodeInfos().Get(allNodes[i].Name)
 					p.Filter(ctx, state, tt.pod, n)
 				}
-				p.parallelizer.Until(ctx, len(allNodes), filterNode)
+				p.parallelizer.Until(ctx, len(allNodes), filterNode, "")
 			}
 		})
 		b.Run(tt.name+"/Clone", func(b *testing.B) {

pkg/scheduler/framework/plugins/podtopologyspread/scoring.go

@@ -183,7 +183,7 @@ func (pl *PodTopologySpread) PreScore(
 			atomic.AddInt64(tpCount, int64(count))
 		}
 	}
-	pl.parallelizer.Until(ctx, len(allNodes), processAllNode)
+	pl.parallelizer.Until(ctx, len(allNodes), processAllNode, pl.Name())
 
 	cycleState.Write(preScoreStateKey, state)
 	return nil

pkg/scheduler/framework/plugins/podtopologyspread/scoring_test.go

@@ -1408,7 +1408,7 @@ func BenchmarkTestDefaultEvenPodsSpreadPriority(b *testing.B) {
 					score, _ := p.Score(ctx, state, pod, n.Name)
 					gotList[i] = framework.NodeScore{Name: n.Name, Score: score}
 				}
-				p.parallelizer.Until(ctx, len(filteredNodes), scoreNode)
+				p.parallelizer.Until(ctx, len(filteredNodes), scoreNode, "")
 				status = p.NormalizeScore(ctx, state, pod, gotList)
 				if !status.IsSuccess() {
 					b.Fatal(status)

pkg/scheduler/framework/plugins/selectorspread/selector_spread_perf_test.go

@@ -90,7 +90,7 @@ func BenchmarkTestSelectorSpreadPriority(b *testing.B) {
 					gotList[i] = framework.NodeScore{Name: n.Name, Score: score}
 				}
 				parallelizer := parallelize.NewParallelizer(parallelize.DefaultParallelism)
-				parallelizer.Until(ctx, len(filteredNodes), scoreNode)
+				parallelizer.Until(ctx, len(filteredNodes), scoreNode, "")
 				status = plugin.NormalizeScore(ctx, state, pod, gotList)
 				if !status.IsSuccess() {
 					b.Fatal(status)

pkg/scheduler/framework/preemption/preemption.go

@@ -611,6 +611,6 @@ func (ev *Evaluator) DryRunPreemption(ctx context.Context, pod *v1.Pod, potentia
 		nodeStatuses[nodeInfoCopy.Node().Name] = status
 		statusesLock.Unlock()
 	}
-	fh.Parallelizer().Until(ctx, len(potentialNodes), checkNode)
+	fh.Parallelizer().Until(ctx, len(potentialNodes), checkNode, ev.PluginName)
 	return append(nonViolatingCandidates.get(), violatingCandidates.get()...), nodeStatuses, utilerrors.NewAggregate(errs)
 }

pkg/scheduler/framework/runtime/framework.go

@@ -925,7 +925,7 @@ func (f *frameworkImpl) RunScorePlugins(ctx context.Context, state *framework.Cy
 				Score: s,
 			}
 		}
-	})
+	}, score)
 	if err := errCh.ReceiveError(); err != nil {
 		return nil, framework.AsStatus(fmt.Errorf("running Score plugins: %w", err))
 	}
@@ -943,7 +943,7 @@ func (f *frameworkImpl) RunScorePlugins(ctx context.Context, state *framework.Cy
 			errCh.SendErrorWithCancel(err, cancel)
 			return
 		}
-	})
+	}, score)
 	if err := errCh.ReceiveError(); err != nil {
 		return nil, framework.AsStatus(fmt.Errorf("running Normalize on Score plugins: %w", err))
 	}
@@ -964,7 +964,7 @@ func (f *frameworkImpl) RunScorePlugins(ctx context.Context, state *framework.Cy
 			}
 			nodeScoreList[i].Score = nodeScore.Score * int64(weight)
 		}
-	})
+	}, score)
 	if err := errCh.ReceiveError(); err != nil {
 		return nil, framework.AsStatus(fmt.Errorf("applying score defaultWeights on Score plugins: %w", err))
 	}

pkg/scheduler/metrics/metrics.go

@@ -95,14 +95,23 @@ var (
 			Help:           "Number of pending pods, by the queue type. 'active' means number of pods in activeQ; 'backoff' means number of pods in backoffQ; 'unschedulable' means number of pods in unschedulablePods.",
 			StabilityLevel: metrics.STABLE,
 		}, []string{"queue"})
+	// SchedulerGoroutines isn't called in some parts where goroutines start.
+	// Goroutines metric replaces SchedulerGoroutines metric. Goroutine metric tracks all goroutines.
 	SchedulerGoroutines = metrics.NewGaugeVec(
 		&metrics.GaugeOpts{
 			Subsystem:         SchedulerSubsystem,
+			DeprecatedVersion: "1.26.0",
 			Name:              "scheduler_goroutines",
-			Help:           "Number of running goroutines split by the work they do such as binding.",
+			Help:              "Number of running goroutines split by the work they do such as binding. This metric is replaced by the \"goroutines\" metric.",
 			StabilityLevel:    metrics.ALPHA,
 		}, []string{"work"})
-
+	Goroutines = metrics.NewGaugeVec(
+		&metrics.GaugeOpts{
+			Subsystem:      SchedulerSubsystem,
+			Name:           "goroutines",
+			Help:           "Number of running goroutines split by the work they do such as binding.",
+			StabilityLevel: metrics.ALPHA,
+		}, []string{"operation"})
 	PodSchedulingDuration = metrics.NewHistogramVec(
 		&metrics.HistogramOpts{
 			Subsystem: SchedulerSubsystem,
@@ -194,6 +203,7 @@ var (
 		PluginExecutionDuration,
 		SchedulerQueueIncomingPods,
 		SchedulerGoroutines,
+		Goroutines,
 		PermitWaitDuration,
 		CacheSize,
 		unschedulableReasons,

pkg/scheduler/schedule_one.go

@@ -103,6 +103,8 @@ func (sched *Scheduler) scheduleOne(ctx context.Context) {
 
 		metrics.SchedulerGoroutines.WithLabelValues(metrics.Binding).Inc()
 		defer metrics.SchedulerGoroutines.WithLabelValues(metrics.Binding).Dec()
+		metrics.Goroutines.WithLabelValues(metrics.Binding).Inc()
+		defer metrics.Goroutines.WithLabelValues(metrics.Binding).Dec()
 
 		sched.bindingCycle(bindingCycleCtx, state, fwk, scheduleResult, assumedPodInfo, podsToActivate, start)
 	}()
@@ -563,7 +565,7 @@ func (sched *Scheduler) findNodesThatPassFilters(
 
 	// Stops searching for more nodes once the configured number of feasible nodes
 	// are found.
-	fwk.Parallelizer().Until(ctx, numAllNodes, checkNode)
+	fwk.Parallelizer().Until(ctx, numAllNodes, checkNode, frameworkruntime.Filter)
 	feasibleNodes = feasibleNodes[:feasibleNodesLen]
 	if err := errCh.ReceiveError(); err != nil {
 		statusCode = framework.Error
@@ -718,8 +720,10 @@ func prioritizeNodes(
 			wg.Add(1)
 			go func(extIndex int) {
 				metrics.SchedulerGoroutines.WithLabelValues(metrics.PrioritizingExtender).Inc()
+				metrics.Goroutines.WithLabelValues(metrics.PrioritizingExtender).Inc()
 				defer func() {
 					metrics.SchedulerGoroutines.WithLabelValues(metrics.PrioritizingExtender).Dec()
+					metrics.Goroutines.WithLabelValues(metrics.PrioritizingExtender).Dec()
 					wg.Done()
 				}()
 				prioritizedList, weight, err := extenders[extIndex].Prioritize(pod, nodes)