Automations → Monitoring shows a window of 7, 14, 30 or 90 days.
Cards: Total Executions, Completed, Failed, Running, Skipped, Dead letters and Avg Duration. The success rate excludes skipped runs, but its base still includes running and waiting ones. The table lists the 50 busiest workflows with runs, health and last result; clicking a row opens the editor.
Caveats
- Dead letters count failures no error branch handled. They are not limited to the window, cannot be listed or closed, and a successful "retry from failed step" does not reduce them.
- Avg Duration includes skipped rows (about zero) and resumed parents, whose duration includes the whole wait.
- Per-workflow health divides by all runs, skipped included, so it disagrees with the overall rate.
- Runs the engine reaped for running too long send no alert.