@Marvin If I have an automation setup to act on flow runs ending in a Failed or Crashed state, will that automation fire on flow runs ending in a TimedOut state?

✅ thought for 145.9 seconds

well failed and crashed are distinct state types, but TimedOut is a named state (ie just a semantically useful wrapper around Failed) so what is your automation definition?

No — not by default. If your automation is listening for: -

prefect.flow-run.Failed

-

prefect.flow-run.Crashed

it will not fire for a flow run that ends in

TimedOut

. Why: -

TimedOut

has a

FAILED

state type - but its state name is

TimedOut

- automation events are emitted from the state name, so the event is

prefect.flow-run.TimedOut

So if you want to catch timeouts too, include it explicitly:

yaml
expect:
  - prefect.flow-run.Failed
  - prefect.flow-run.Crashed
  - prefect.flow-run.TimedOut

Relevant docs: - Prefect states documentation If you want, I can also help you translate that into the exact automation UI config or JSON/YAML you're using.

Thanks, that's what I figured, just making sure I understand the why. Our notification trigger is:

AutomationTrigger(
                event=AutomationTriggerEvent(
                    posture='Reactive',
                    threshold=1,
                    within=0,
                    expect=['prefect.flow-run.Failed', 'prefect.flow-run.Crashed'],
                    match=dumps(
                        {
                            'prefect.resource.id': 'prefect.flow-run.*',
                        }
                    ),
                    match_related=dumps(
                        {
                            'prefect.resource.id': f'prefect.tag.{SLACK_NOTIFICATION_TAG_PREFIX}:*',
                            'prefect.resource.role': 'tag',
                        }
                    ),
                )
            ),