-
Notifications
You must be signed in to change notification settings - Fork 11
Open
Description
The plan for #123 is to aggregate / merge metrics for all tasks together so we have "per stage" metrics. Each node may have metrics in the form of output_rows = X, output_rows = Y, output_rows = Z
where X, Y, Z are metrics for each task. The aggregation output_rows = X + Y + Z
is shown in the explain analyze output.
Unfortunately, this hides metrics per task, which would be useful during debugging.
There's some options to make metrics inspectable
- add labels to metrics so they are tagged with task id
- make EXPLAIN VERBOSE display metrics per task
- add a debug log line which makes each task emit a log of its local plan with metrics
Metadata
Metadata
Assignees
Labels
No labels