collect started metric

Author: Trojan295

internal/controller/controller.go

@@ -156,7 +156,9 @@ func (s *Controller) handleActions(ctx context.Context, clusterActions []*castai
 
 			var err error
 
+			metrics.ActionStarted(action.GetType())
 			startTime := time.Now()
+
 			handleErr := s.handleAction(ctx, action)
 			if errors.Is(handleErr, context.Canceled) {
 				// Action should be handled again on context canceled errors.

internal/metrics/custom_metrics.go

@@ -7,6 +7,7 @@ import (
 	"github.com/prometheus/client_golang/prometheus"
 )
 
+// actionStartedCounter tracks actions started by the cluster controller.
 var actionStartedCounter = prometheus.NewCounterVec(
 	prometheus.CounterOpts{
 		Name: "action_started_total",
@@ -24,6 +25,11 @@ var actionExecutedCounter = prometheus.NewCounterVec(
 	[]string{"success", "type"},
 )
 
+// actionExecutedDuration tracks the duration of actions executed by the cluster controller.
+// Summary is used instead of histogram to use less series.
+// We cannot run aggregations on Prometheus side with Summaries,
+// but there is only a single active cluster controller pod,
+// so this might not be a problem.
 var actionExecutedDuration = prometheus.NewSummaryVec(
 	prometheus.SummaryOpts{
 		Name: "action_executed_duration_seconds",

internal/metrics/register.go

@@ -6,8 +6,8 @@ import (
 
 func RegisterCustomMetrics() {
 	registry.MustRegister(
-		actionCounter,
+		actionStartedCounter,
+		actionExecutedCounter,
 		actionExecutedDuration,
-		actionExecutedDurationSummary,
 	)
 }