BasPH commented on PR #39298:
URL: https://github.com/apache/airflow/pull/39298#issuecomment-2089948475

   Okay I reverted a few logging statements so that now leaves an info log when 
`AIRFLOW__CORE__PARALLELISM` is reached, which should help users understand why 
Airflow isn't running any more tasks.
   
   @hussein-awala Regarding:
   > It would be useless to warn the user when the pool limit or dag/run/task 
max concurrency are reached, because these limits are usually set according to 
the capacity of an external system
   
   Disagree with this statement. Airflow has lots of knobs to tweak for 
parallelism which can be a blessing and a curse. You might have a perfect setup 
with OTEL integration, dashboards, and alerts, but I'm sure that knowledge and 
tooling wasn't developed overnight. My experience is that Airflow's default 
settings quite often limit the user before they start reaching hardware limits. 
This is a frustrating problem because as of today visibility (via UI/logs) on 
such limits is low. I'll investigate alternative options such as dashboards, 
but stating that it's "useless to warn the user" is not my experience since 9 
out of 10 times I see people hitting a default limit instead of a user-defined 
limit.


-- 
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