metrics

Documentation

Index

• Constants
• Variables
• func ExecuteErrorToLabel(err error) string
• func RetLabel(err error) string

Constants

const (
	LabelSession  = "session"
	LabelDomain   = "domain"
	LabelDDL      = "ddl"
	LabelGCWorker = "gcworker"
)

metrics labels.

const (
	LblUnretryable = "unretryable"
	LblReachMax    = "reach_max"
	LblOK          = "ok"
	LblError       = "error"
	LblRollback    = "rollback"
	LblType        = "type"
	LblResult      = "result"
)

Label constants.

Variables

var (
	JobsGauge = prometheus.NewGaugeVec(
		prometheus.GaugeOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "waiting_jobs",
			Help:      "Gauge of jobs.",
		}, []string{LblType})

	HandleJobHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "handle_job_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of handle jobs",
			Buckets:   prometheus.ExponentialBuckets(0.01, 2, 20),
		}, []string{LblType, LblResult})

	BatchAddIdxHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "batch_add_idx_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of batch handle data",
			Buckets:   prometheus.ExponentialBuckets(0.001, 2, 20),
		})

	SyncerInit    = "init"
	SyncerRestart = "restart"
	SyncerClear   = "clear"

	DeploySyncerHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "deploy_syncer_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of deploy syncer",
			Buckets:   prometheus.ExponentialBuckets(0.01, 2, 20),
		}, []string{LblType, LblResult})

	UpdateSelfVersionHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "update_self_ver_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of update self version",
			Buckets:   prometheus.ExponentialBuckets(0.01, 2, 20),
		}, []string{LblResult})

	OwnerUpdateGlobalVersion   = "update_global_version"
	OwnerGetGlobalVersion      = "get_global_version"
	OwnerCheckAllVersions      = "check_all_versions"
	OwnerHandleSyncerHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "owner_handle_syncer_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of handle syncer",
			Buckets:   prometheus.ExponentialBuckets(0.01, 2, 20),
		}, []string{LblType, LblResult})

	// Metrics for ddl_worker.go.
	WorkerAddDDLJob         = "add_job"
	WorkerFinishDDLJob      = "finish_job"
	WorkerWaitSchemaChanged = "wait_schema_changed"
	DDLWorkerHistogram      = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "worker_operation_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of ddl worker operations",
			Buckets:   prometheus.ExponentialBuckets(0.001, 2, 20),
		}, []string{LblType, LblResult})

	CreateDDL       = "create_ddl"
	CreateDDLWorker = "create_ddl_worker"
	IsDDLOwner      = "is_ddl_owner"
	DDLCounter      = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "ddl",
			Name:      "worker_operation_total",
			Help:      "Counter of creating ddl/worker and isowner.",
		}, []string{LblType})
)

Metrics for the DDL package.

var (
	DistSQLQueryHistgram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "distsql",
			Name:      "handle_query_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of handled queries.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 13),
		}, []string{LblType})

	DistSQLScanKeysPartialHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "distsql",
			Name:      "scan_keys_partial_num",
			Help:      "number of scanned keys for each partial result.",
		},
	)
	DistSQLScanKeysHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "distsql",
			Name:      "scan_keys_num",
			Help:      "number of scanned keys for each query.",
		},
	)
	DistSQLPartialCountHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "distsql",
			Name:      "partial_num",
			Help:      "number of partial results for each query.",
		},
	)
)

distsql metrics.

var (
	// LoadSchemaCounter records the counter of load schema.
	LoadSchemaCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "domain",
			Name:      "load_schema_total",
			Help:      "Counter of load schema",
		}, []string{LblType})

	// LoadSchemaDuration records the duration of load schema.
	LoadSchemaDuration = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "domain",
			Name:      "load_schema_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) in load schema.",
			Buckets:   prometheus.ExponentialBuckets(0.001, 2, 15),
		})

	// LoadPrivilegeCounter records the counter of load privilege.
	LoadPrivilegeCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "domain",
			Name:      "load_privilege_total",
			Help:      "Counter of load privilege",
		}, []string{LblType})
)

var (
	// ExecutorCounter records the number of expensive executors.
	ExecutorCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "executor",
			Name:      "expensive_total",
			Help:      "Counter of Expensive Executors.",
		}, []string{LblType},
	)

	// StmtNodeCounter records the number of statement with the same type.
	StmtNodeCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "executor",
			Name:      "statement_total",
			Help:      "Counter of StmtNode.",
		}, []string{LblType})
)

var (
	GlobalAutoID      = "global"
	TableAutoIDAlloc  = "alloc"
	TableAutoIDRebase = "rebase"
	AutoIDHistogram   = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "autoid",
			Name:      "operation_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of handled autoid.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 22),
		}, []string{LblType, LblResult})

	GetSchemaDiff    = "get_schema_diff"
	SetSchemaDiff    = "set_schema_diff"
	GetDDLJob        = "get_ddl_job"
	UpdateDDLJob     = "update_ddl_job"
	GetHistoryDDLJob = "get_history_ddl_job"

	MetaHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "meta",
			Name:      "operation_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of tidb meta data operations.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 22),
		}, []string{LblType, LblResult})
)

Metrics

var (
	NewSessionHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "owner",
			Name:      "new_session_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of new session.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 22),
		}, []string{LblType, LblResult})

	Cancelled         = "cancelled"
	Deleted           = "deleted"
	SessionDone       = "session_done"
	CtxDone           = "context_done"
	WatchOwnerCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "owner",
			Name:      "watch_owner_total",
			Help:      "Counter of watch owner.",
		}, []string{LblType, LblResult})

	NoLongerOwner        = "no_longer_owner"
	CampaignOwnerCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "owner",
			Name:      "campaign_owner_total",
			Help:      "Counter of campaign owner.",
		}, []string{LblType, LblResult})
)

Metrics

var (
	QueryDurationHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "server",
			Name:      "handle_query_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of handled queries.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 22),
		})

	QueryTotalCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "server",
			Name:      "query_total",
			Help:      "Counter of queries.",
		}, []string{LblType, LblResult})

	ConnGauge = prometheus.NewGauge(
		prometheus.GaugeOpts{
			Namespace: "tidb",
			Subsystem: "server",
			Name:      "connections",
			Help:      "Number of connections.",
		})

	ExecuteErrorCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "server",
			Name:      "execute_error_total",
			Help:      "Counter of execute errors.",
		}, []string{LblType})

	CriticalErrorCounter = prometheus.NewCounter(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "server",
			Name:      "critical_error_total",
			Help:      "Counter of critical errors.",
		})

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

	TimeJumpBackCounter = prometheus.NewCounter(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "monitor",
			Name:      "time_jump_back_total",
			Help:      "Counter of system time jumps backward.",
		})

	KeepAliveCounter = prometheus.NewCounter(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "monitor",
			Name:      "keep_alive_total",
			Help:      "Counter of TiDB keep alive.",
		})
)

Metrics

var (
	SessionExecuteParseDuration = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "parse_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) in parse SQL.",
			Buckets:   prometheus.LinearBuckets(0.00004, 0.00001, 13),
		})
	SessionExecuteCompileDuration = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "compile_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) in query optimize.",
			Buckets:   prometheus.LinearBuckets(0.00004, 0.00001, 13),
		})
	SessionExecuteRunDuration = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "execute_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) in running executor.",
			Buckets:   prometheus.ExponentialBuckets(0.0001, 2, 13),
		})
	SchemaLeaseErrorCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "schema_lease_error_total",
			Help:      "Counter of schema lease error",
		}, []string{LblType})
	SessionRetry = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "retry_num",
			Help:      "Bucketed histogram of session retry count.",
			Buckets:   prometheus.LinearBuckets(0, 1, 10),
		})
	SessionRetryErrorCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "retry_error_total",
			Help:      "Counter of session retry error.",
		}, []string{LblType})
	TransactionCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "transaction_total",
			Help:      "Counter of transactions.",
		}, []string{LblType})

	SessionRestrictedSQLCounter = prometheus.NewCounter(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "restricted_sql_total",
			Help:      "Counter of internal restricted sql.",
		})

	StatementPerTransaction = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "transaction_statement_num",
			Help:      "Buckated histogram of statements count in each transaction.",
			Buckets:   prometheus.ExponentialBuckets(1, 2, 12),
		}, []string{LblType})

	TransactionDuration = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "session",
			Name:      "transaction_duration_seconds",
			Help:      "Bucketed histogram of a transaction execution duration, including retry.",
			Buckets:   prometheus.ExponentialBuckets(0.001, 2, 16),
		}, []string{LblType})
)

Session metrics.

var (
	AutoAnalyzeHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "statistics",
			Name:      "auto_analyze_duration_seconds",
			Help:      "Bucketed histogram of processing time (s) of auto analyze.",
			Buckets:   prometheus.ExponentialBuckets(0.01, 2, 20),
		})

	AutoAnalyzeCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "statistics",
			Name:      "auto_analyze_total",
			Help:      "Counter of auto analyze.",
		}, []string{LblType})

	StatsInaccuracyRate = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "statistics",
			Name:      "stats_inaccuracy_rate",
			Help:      "Bucketed histogram of stats inaccuracy rate.",
			Buckets:   prometheus.ExponentialBuckets(0.01, 2, 14),
		})

	PseudoEstimation = prometheus.NewCounter(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "statistics",
			Name:      "pseudo_estimation_total",
			Help:      "Counter of pseudo estimation caused by outdated stats.",
		})
)

Stats metrics.

var (
	TiKVTxnCounter = prometheus.NewCounter(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "txn_total",
			Help:      "Counter of created txns.",
		})

	TiKVSnapshotCounter = prometheus.NewCounter(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "snapshot_total",
			Help:      "Counter of snapshots.",
		})

	TiKVTxnCmdCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "txn_cmd_total",
			Help:      "Counter of txn commands.",
		}, []string{LblType})

	TiKVTxnCmdHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "txn_cmd_duration_seconds",
			Help:      "Bucketed histogram of processing time of txn cmds.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 20),
		}, []string{LblType})

	TiKVBackoffCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "backoff_total",
			Help:      "Counter of backoff.",
		}, []string{LblType})

	TiKVBackoffHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "backoff_seconds",
			Help:      "total backoff seconds of a single backoffer.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 20),
		})

	TiKVConnPoolHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "get_conn_seconds",
			Help:      "Bucketed histogram of taking conn from conn pool.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 20),
		}, []string{LblType})

	TiKVSendReqHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "request_seconds",
			Help:      "Bucketed histogram of sending request duration.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 20),
		}, []string{LblType, "store"})

	TiKVCoprocessorCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "cop_actions_total",
			Help:      "Counter of coprocessor actions.",
		}, []string{LblType})

	TiKVCoprocessorHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "cop_duration_seconds",
			Help:      "Run duration of a single coprocessor task, includes backoff time.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 20),
		})

	TiKVLockResolverCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "lock_resolver_actions_total",
			Help:      "Counter of lock resolver actions.",
		}, []string{LblType})

	TiKVRegionErrorCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "region_err_total",
			Help:      "Counter of region errors.",
		}, []string{LblType})

	TiKVTxnWriteKVCountHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "txn_write_kv_num",
			Help:      "Count of kv pairs to write in a transaction.",
			Buckets:   prometheus.ExponentialBuckets(1, 2, 21),
		})

	TiKVTxnWriteSizeHistogram = prometheus.NewHistogram(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "txn_write_size_bytes",
			Help:      "Size of kv pairs to write in a transaction.",
			Buckets:   prometheus.ExponentialBuckets(1, 2, 21),
		})

	TiKVRawkvCmdHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "rawkv_cmd_seconds",
			Help:      "Bucketed histogram of processing time of rawkv cmds.",
			Buckets:   prometheus.ExponentialBuckets(0.0005, 2, 20),
		}, []string{LblType})

	TiKVRawkvSizeHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "rawkv_kv_size_bytes",
			Help:      "Size of key/value to put, in bytes.",
			Buckets:   prometheus.ExponentialBuckets(1, 2, 21),
		}, []string{LblType})

	TiKVTxnRegionsNumHistogram = prometheus.NewHistogramVec(
		prometheus.HistogramOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "txn_regions_num",
			Help:      "Number of regions in a transaction.",
			Buckets:   prometheus.ExponentialBuckets(1, 2, 20),
		}, []string{LblType})

	TiKVLoadSafepointCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "load_safepoint_total",
			Help:      "Counter of load safepoint.",
		}, []string{LblType})

	TiKVSecondaryLockCleanupFailureCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "lock_cleanup_task_total",
			Help:      "failure statistic of secondary lock cleanup task.",
		}, []string{LblType})

	TiKVRegionCacheCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "tikvclient",
			Name:      "region_cache_operations_total",
			Help:      "Counter of region cache.",
		}, []string{LblType, LblResult})
)

TiKVClient metrics.

var (
	// PanicCounter measures the count of panics.
	PanicCounter = prometheus.NewCounterVec(
		prometheus.CounterOpts{
			Namespace: "tidb",
			Subsystem: "server",
			Name:      "panic_total",
			Help:      "Counter of panic.",
		}, []string{LblType})
)

Functions

func ExecuteErrorToLabel

func ExecuteErrorToLabel(err error) string

ExecuteErrorToLabel converts an execute error to label.

func RetLabel

func RetLabel(err error) string

RetLabel returns "ok" when err == nil and "err" when err != nil. This could be useful when you need to observe the operation result.