add metric to track started actions

Author: Trojan295

internal/castai/client_test.go

@@ -250,7 +250,7 @@ func TestConvertPrometheusMetricFamilies(t *testing.T) {
 		metricName := "test_histogram"
 		sampleCount := uint64(100)
 		sampleSum := 250.5
-		
+
 		// Create histogram buckets
 		buckets := []*dto.Bucket{
 			{
@@ -334,20 +334,20 @@ func TestConvertPrometheusMetricFamilies(t *testing.T) {
 
 		// Verify bucket timeseries (4 explicit buckets + 1 +Inf bucket)
 		r.Len(bucketTimeseries, 5)
-		
+
 		// Verify each expected bucket exists
 		expectedBuckets := map[string]float64{
-			"1.000000": 10,
-			"5.000000": 30,
+			"1.000000":  10,
+			"5.000000":  30,
 			"10.000000": 80,
 			"25.000000": 100,
-			"+Inf": 100,
+			"+Inf":      100,
 		}
 
 		foundBuckets := make(map[string]bool)
 		for _, ts := range bucketTimeseries {
 			assertLabelPresent(t, ts.Labels, "__name__", metricName+"_bucket")
-			
+
 			// Find the 'le' label value
 			var leValue string
 			for _, label := range ts.Labels {
@@ -356,38 +356,42 @@ func TestConvertPrometheusMetricFamilies(t *testing.T) {
 					break
 				}
 			}
-			
+
 			r.NotEmpty(leValue, "Bucket timeseries should have 'le' label")
-			
+
 			expectedCount, exists := expectedBuckets[leValue]
 			r.True(exists, "Unexpected bucket with le=%s", leValue)
 			r.Equal(expectedCount, ts.Samples[0].Value)
 			foundBuckets[leValue] = true
 		}
-		
+
 		// Ensure all expected buckets were found
 		r.Len(foundBuckets, len(expectedBuckets))
 
 		// Verify _sum timeseries
 		r.NotNil(sumTimeseries)
 		assertLabelPresent(t, sumTimeseries.Labels, "__name__", metricName+"_sum")
 		r.Equal(sampleSum, sumTimeseries.Samples[0].Value)
-		
+
 		// Verify _sum doesn't have 'le' label
 		for _, label := range sumTimeseries.Labels {
 			r.NotEqual("le", label.Name, "_sum should not have 'le' label")
 		}
 
-		// Verify _count timeseries  
+		// Verify _count timeseries
 		r.NotNil(countTimeseries)
 		assertLabelPresent(t, countTimeseries.Labels, "__name__", metricName+"_count")
 		r.Equal(float64(sampleCount), countTimeseries.Samples[0].Value)
-		
+
 		// Verify _count doesn't have 'le' label
 		for _, label := range countTimeseries.Labels {
 			r.NotEqual("le", label.Name, "_count should not have 'le' label")
 		}
 	})
+
+	t.Run("summary", func(t *testing.T) {
+		// TODO: unit test for summary metrics
+	})
 }
 
 func assertLabelPresent(t *testing.T, labels []PrometheusLabel, name, value string) {

internal/metrics/custom_metrics.go

@@ -7,40 +7,41 @@ import (
 	"github.com/prometheus/client_golang/prometheus"
 )
 
-// actionCounter tracks actions executed by the cluster controller.
-var actionCounter = prometheus.NewCounterVec(
+var actionStartedCounter = prometheus.NewCounterVec(
 	prometheus.CounterOpts{
-		Name: "action_executed_total",
-		Help: "Count of successful and unsuccessful actions executed by type.",
+		Name: "action_started_total",
+		Help: "Count of actions started by type.",
 	},
-	[]string{"success", "type"},
+	[]string{"type"},
 )
 
-// actionDuration tracks the duration of actions executed by the cluster controller.
-var actionExecutedDuration = prometheus.NewHistogramVec(
-	prometheus.HistogramOpts{
-		Name:    "action_executed_duration_seconds",
-		Help:    "Duration of action handle execution in seconds.",
-		Buckets: prometheus.ExponentialBucketsRange(0.01, 30, 7),
+// actionExecutedCounter tracks actions executed by the cluster controller.
+var actionExecutedCounter = prometheus.NewCounterVec(
+	prometheus.CounterOpts{
+		Name: "action_executed_total",
+		Help: "Count of successful and unsuccessful actions executed by type.",
 	},
-	[]string{"type"},
+	[]string{"success", "type"},
 )
 
-var actionExecutedDurationSummary = prometheus.NewSummaryVec(
+var actionExecutedDuration = prometheus.NewSummaryVec(
 	prometheus.SummaryOpts{
-		Name: "action_executed_duration_summary_seconds",
-		Help: "Duration of action handle execution in seconds.",
+		Name: "action_executed_duration_seconds",
+		Help: "Duration of actions executed by type.",
 		Objectives: map[float64]float64{
-			0.5:  0.05,  // 50th percentile (median) with 5% error
-			0.9:  0.01,  // 90th percentile with 1% error
-			0.99: 0.001, // 99th percentile with 0.1% error
+			0.5:  0.05,
+			0.9:  0.01,
+			0.99: 0.001,
 		},
 	},
 	[]string{"type"},
 )
 
+func ActionStarted(actionType string) {
+	actionStartedCounter.With(prometheus.Labels{"type": actionType}).Inc()
+}
+
 func ActionFinished(actionType string, success bool, duration time.Duration) {
-	actionCounter.With(prometheus.Labels{"success": strconv.FormatBool(success), "type": actionType}).Inc()
+	actionExecutedCounter.With(prometheus.Labels{"success": strconv.FormatBool(success), "type": actionType}).Inc()
 	actionExecutedDuration.With(prometheus.Labels{"type": actionType}).Observe(duration.Seconds())
-	actionExecutedDurationSummary.With(prometheus.Labels{"type": actionType}).Observe(duration.Seconds())
 }