Noora health long job timeout failure

[Ishankoradia]

One job of noora health configured has run multiple airbyte syncs (~ 20), which run sequentially one after the other.

Some observations from the recent failures

• Failure can occure at any random step. Usually after a few steps are over.
• The logs in prefect show Task run failed with exception: ReadTimeout('') - Retries are exhausted.
• The step at which the job fails, seems to have executed successfully in airbyte. Saw this by comparing prefect and airbyte timestamps. Which means, prefect successfully triggered the connection but might have lost the connection to poll on. Found this log message in one of the airbyte workers 2025-08-23 19:47:16,821 [scheduler-executor-thread-6] INFO i.a.c.s.h.SchedulerHandler(createJob):644 - New job created, with id: 44003 . This is the id of the step that failed according to prefect.
• From the error stack, i can see that this line 388 in prefect_airybte is the call where it failed. Maybe that api is timing out.

• Works when retrying.

[Ishankoradia]

Immediate workaround would be to do a retry from Dalgo in the webhook itself.