Skip to main content
PlanetNodes

Monitor, cancel, and retry jobs

Respond correctly to queued, active, failed, and terminal job states.

All guides →

Monitor

Poll the job resource or use the event stream and configured webhook for lifecycle updates. Treat terminal states as authoritative.

Cancel deliberately

Cancellation is appropriate when queued work is no longer needed. An already-running provider may need time to acknowledge the transition.

Retry safely

  • Retry transient provider or infrastructure failures.
  • Do not automatically retry invalid payloads.
  • Use your own idempotency boundary when downstream side effects matter.

Related guides