potiuk commented on pull request #15212:
URL: https://github.com/apache/airflow/pull/15212#issuecomment-813696590


   > Generally, though, I have found that a surprising number of tools will 
happily work without an actual user record, so there is a user expectation here 
when they're firing things into Docker. I think it doesn't make sense to match 
that given Airflow is relatively dependent on user env/environment variables, 
but I'll see if I can make an error message that is nice enough to stop 
duplicate issues like #9963 being opened.
   
   I do not think this is the user expectation. Simply the users are unaware 
this is a problem and it's better to educate them.
   Much more surprising to the users, is that = they find that SOME of their 
dependencies do not work, and then it usually will end up with `airflow` being 
the culprit and issues raised towards the airflow community. I prefer much more 
to fail hard in this case and display an explicit error message rather than 
"pretend nothing happens". The cost of fixing such problem is directly 
proportional to the time it is needed to detect and diagnose the problem. The 
sooner we detect it and the more explicit we get, the less it cost for the user 
and the community to fix it. It's very low cost to detect the problem 
immediately when you run 'airflow'. It is far more costly to fix it when you 
find out after few months of running the DAGs that some of your tools 
misbehaved but you have not noticed it.
   


-- 
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.

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


Reply via email to