Skip to main content
Version: Nightly

FLOW_STATISTICS

The FLOW_STATISTICS table provides runtime statistics for flows. It lists every flow in the current catalog, including flows that have not reported statistics yet.

DESC TABLE INFORMATION_SCHEMA.FLOW_STATISTICS;
+---------------------+----------------------+------+------+---------+---------------+
| Column              | Type                 | Key  | Null | Default | Semantic Type |
+---------------------+----------------------+------+------+---------+---------------+
| flow_id             | UInt32               |      | NO   |         | FIELD         |
| flow_name           | String               |      | NO   |         | FIELD         |
| start_time          | TimestampMillisecond |      | YES  |         | FIELD         |
| last_execution_time | TimestampMillisecond |      | YES  |         | FIELD         |
| uptime_seconds      | Int64                |      | YES  |         | FIELD         |
| state_size          | UInt64               |      | YES  |         | FIELD         |
+---------------------+----------------------+------+------+---------+---------------+

Fields in the FLOW_STATISTICS table are described as follows:

  • flow_id: The ID of the flow.
  • flow_name: The name of the flow.
  • start_time: Timestamp indicating when the Flownode started running the flow.
  • last_execution_time: Timestamp of the flow's most recent execution.
  • uptime_seconds: Seconds elapsed since start_time.
  • state_size: Size of the flow's in-memory state, in bytes.

Statistics are reported by Flownodes through heartbeats. The last four columns are NULL for a flow that has no statistics reported yet, for example immediately after the flow is created or after a Flownode restart.

In a cluster, start_time and uptime_seconds are always NULL: the Flownode state persisted for the cluster does not carry a flow's start time. Both columns are populated in standalone deployments.

SHOW FLOW STATUS returns the same columns.