metrics: Clean up metrics (pingcap#5858)

metrics/ddl.go

@@ -43,9 +43,9 @@ var (
 			Buckets:   prometheus.ExponentialBuckets(0.001, 2, 20),
 		})
 
-	SyncerInit    = "syncer_init"
-	SyncerRestart = "syncer_restart"
-	SyncerClear   = "syncer_clear"
+	SyncerInit    = "init"
+	SyncerRestart = "restart"
+	SyncerClear   = "clear"
 
 	DeploySyncerHistogram = prometheus.NewHistogramVec(
 		prometheus.HistogramOpts{
@@ -78,14 +78,14 @@ var (
 		}, []string{"op", "result_state"})
 
 	// Metrics for ddl_worker.go.
-	WorkerAddDDLJob         = "add_ddl_job"
-	WorkerFinishDDLJob      = "finish_ddl_job"
+	WorkerAddDDLJob         = "add_job"
+	WorkerFinishDDLJob      = "finish_job"
 	WorkerWaitSchemaChanged = "wait_schema_changed"
 	DDLWorkerHistogram      = prometheus.NewHistogramVec(
 		prometheus.HistogramOpts{
 			Namespace: "tidb",
 			Subsystem: "ddl",
-			Name:      "ddl_worker_operation",
+			Name:      "worker_operation_duration_seconds",
 			Help:      "Bucketed histogram of processing time (s) of ddl worker operations",
 			Buckets:   prometheus.ExponentialBuckets(0.001, 2, 20),
 		}, []string{"op", "result_state"})
@@ -97,7 +97,7 @@ var (
 		prometheus.CounterOpts{
 			Namespace: "tidb",
 			Subsystem: "ddl",
-			Name:      "ddl_counter",
+			Name:      "worker_total",
 			Help:      "Counter of creating ddl/worker and isowner.",
 		}, []string{"type"})
 )

metrics/meta.go

@@ -20,13 +20,13 @@ import (
 // Metrics
 var (
 	GlobalAutoID      = "global"
-	TableAutoIDAlloc  = "table"
-	TableAutoIDRebase = "table_rebase"
+	TableAutoIDAlloc  = "alloc"
+	TableAutoIDRebase = "rebase"
 	AutoIDHistogram   = prometheus.NewHistogramVec(
 		prometheus.HistogramOpts{
 			Namespace: "tidb",
-			Subsystem: "meta",
-			Name:      "autoid_alloc",
+			Subsystem: "autoid",
+			Name:      "operation_duration_seconds",
 			Help:      "Bucketed histogram of processing time (s) of handled autoid.",
 			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 22),
 		}, []string{"op", "result_state"})
@@ -41,7 +41,7 @@ var (
 		prometheus.HistogramOpts{
 			Namespace: "tidb",
 			Subsystem: "meta",
-			Name:      "meta_opt",
+			Name:      "operation_duration_seconds",
 			Help:      "Bucketed histogram of processing time (s) of tidb meta data operations.",
 			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 22),
 		}, []string{"op", "result_state"})

metrics/metrics.go

@@ -32,11 +32,11 @@ const (
 	LabelDDL      = "ddl"
 	LabelGCWorker = "gcworker"
 
-	opSucc   = "op_succ"
-	opFailed = "op_failed"
+	opSucc   = "ok"
+	opFailed = "err"
 )
 
-// RetLabel returns "op_succ" when err == nil and "op_failed" when err != nil.
+// RetLabel returns "ok" when err == nil and "err" when err != nil.
 // This could be useful when you need to observe the operation result.
 func RetLabel(err error) string {
 	if err == nil {

metrics/server.go

@@ -64,18 +64,18 @@ var (
 			Help:      "Counter of critical errors.",
 		})
 
-	EventStart        = "server_start"
-	EventGracefulDown = "server_graceful_shutdown"
+	EventStart        = "start"
+	EventGracefulDown = "graceful_shutdown"
 	// Eventkill occurs when the server.Kill() function is called.
-	EventKill = "server_kill"
+	EventKill = "kill"
 	// EventHang occurs when server meet some critical error. It will close the listening port and hang for ever.
-	EventHang          = "server_hang"
-	EventClose         = "server_close"
+	EventHang          = "hang"
+	EventClose         = "close"
 	ServerEventCounter = prometheus.NewCounterVec(
 		prometheus.CounterOpts{
 			Namespace: "tidb",
 			Subsystem: "server",
-			Name:      "server_event",
+			Name:      "event_total",
 			Help:      "Counter of tidb-server event.",
 		}, []string{"type"})
 )