The short answer
Availability: Available to accounts with access to this workspace or feature.
Step-by-step
-
Review reliability metrics
Check worker health, oldest due job, queue lag, last reconciliation and repeated-failure alerts. A separate production worker reports through its database heartbeat.
-
Open a workflow run
Inspect its event, version, status, recorded branch path and node-job timeline.
-
Highlight the path
Use Highlight path on canvas to compare the recorded route with the published workflow layout.
-
Inspect failed work
Review the sanitized error and attempt count. Correct the underlying configuration or provider issue before retrying.
-
Retry or cancel carefully
Retry a failed job only after correction. Cancel only queued or retrying work that should no longer execute.
-
Run reconciliation when needed
Use the rate-limited reconciliation control to recover expired leases and finalize stranded runs. It does not recreate every missing upstream domain event.
Was this guide helpful?
Send feedback about this guide