CyberDuck79 commented on issue #30264:
URL: https://github.com/apache/airflow/issues/30264#issuecomment-2551169610

   Well, after experimentation, it seems there is confusion about callback 
configurations.
   There is DAG level callbacks: 
https://github.com/apache/airflow/discussions/38897
   But we were also confused by the configurations:
   - In the DAG config in `default_args` (our old way of configuring callbacks) 
-> it configures the callbacks at the task level for all tasks.
   - In the DAG config directly (`on_failure_callback`, for example) -> 
callback triggered at dag level by dag state (so in case of dag run timeout, 
failed state -> failure callback)
   The second configuration would allow the callback to be triggered even if 
the task process is killed.
   
   For the question of the skipped status:
   Maybe the skipped status is not very clear because the task could have 
already started, but this information is included in the skipped callback so 
the DAG run timeout case can be handled. It seems that in Airflow, the meaning 
of the skipped state includes unfinished tasks.
   An unfinished state could be more clear, but is it worse to add a new state 
to handle?
   
   In regard to these discoveries, I think the issue should be closed because 
there is nothing that cannot be handled, and a conversation about a possible 
new "unfinished" state for those who find it interesting should be initiated.
   I hope my comments could help to clarify the confusions that I read in this 
conversation.


-- 
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