Anurag-S1ngh commented on PR #77087:
URL: https://github.com/apache/airflow/pull/77087#issuecomment-6998233285

   On the two-minute window: keeping it. A watcher that stays up for 120s 
worked; I count that as a recovery, so a later crash starts at one failure and 
waits the base 5s. Making the window grow with the backoff (up to the 300s cap) 
would leave a watcher that is healthy again stuck as failing for minutes.
   
   On `failing`: keeping one meaning, failed since last recovery, whether the 
trigger is waiting for its backoff or re-running after one. Both are unhealthy 
and want attention. The API already exposes `consecutive_failures` and the 
failure timestamps, so how stale the failure is visible. Having `failing` mean 
only waiting-to-retry would need a backoff deadline persisted on the row and a 
change to the status expression.


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to