add gc metrics

Signed-off-by: yeya24 <[email protected]>

docs/user_guide/metrics.md

@@ -26,6 +26,8 @@ dragonfly_supernode_cdn_download_failed_total          |
 dragonfly_supernode_pieces_downloaded_size_bytes_total |                                        | counter   | Total size of pieces downloaded from supernode in bytes.
 dragonfly_supernode_gc_peers_total                     |                                        | counter   | Total number of peers that have been garbage collected.
 dragonfly_supernode_gc_tasks_total                     |                                        | counter   | Total number of tasks that have been garbage collected.
+dragonfly_supernode_gc_disks_total                     |                                        | counter   | Total number of garbage collecting the task data in disks.
+dragonfly_supernode_last_gc_disks_timestamp_seconds    |                                        | gauge     | Timestamp of the last disk gc.
 
 ## Dfdaemon
 

supernode/config/config.go

@@ -237,8 +237,8 @@ type BaseProperties struct {
 	// IntervalThreshold is the threshold of the interval at which the task file is accessed.
 	IntervalThreshold time.Duration `yaml:"IntervalThreshold"`
 
-	// CleanRatio the ratio to clean the disk and based on 10.
-	// And the value of CleanRatio should be [1-10].
+	// CleanRatio is the ratio to clean the disk and it is based on 10.
+	// It means the value of CleanRatio should be [1-10].
 	//
 	// default: 1
 	CleanRatio int

supernode/daemon/mgr/gc/gc_disk.go

@@ -69,6 +69,8 @@ func (gcm *Manager) deleteTaskDisk(ctx context.Context, gcTaskIDs []string) {
 		util.ReleaseLock(taskID, false)
 		count++
 	}
+	gcm.metrics.gcDisksCount.WithLabelValues().Add(float64(count))
+	gcm.metrics.lastGCDisksTime.WithLabelValues().SetToCurrentTime()
 
 	logrus.Debugf("gc disk: success to gc task count(%d), remainder count(%d)", count, len(gcTaskIDs)-count)
 }

supernode/daemon/mgr/gc/gc_manager.go

@@ -31,16 +31,25 @@ import (
 var _ mgr.GCMgr = &Manager{}
 
 type metrics struct {
-	gcTasksCount *prometheus.CounterVec
-	gcPeersCount *prometheus.CounterVec
+	gcTasksCount    *prometheus.CounterVec
+	gcPeersCount    *prometheus.CounterVec
+	gcDisksCount    *prometheus.CounterVec
+	lastGCDisksTime *prometheus.GaugeVec
 }
 
 func newMetrics(register prometheus.Registerer) *metrics {
 	return &metrics{
 		gcTasksCount: metricsutils.NewCounter(config.SubsystemSupernode, "gc_tasks_total",
 			"Total number of tasks that have been garbage collected", []string{}, register),
+
 		gcPeersCount: metricsutils.NewCounter(config.SubsystemSupernode, "gc_peers_total",
 			"Total number of peers that have been garbage collected", []string{}, register),
+
+		gcDisksCount: metricsutils.NewCounter(config.SubsystemSupernode, "gc_disks_total",
+			"Total number of garbage collecting the task data in disks", []string{}, register),
+
+		lastGCDisksTime: metricsutils.NewGauge(config.SubsystemSupernode, "last_gc_disks_timestamp_seconds",
+			"Timestamp of the last disk gc", []string{}, register),
 	}
 }