potiuk commented on PR #59885: URL: https://github.com/apache/airflow/pull/59885#issuecomment-3697530023
Also - the **high** percentage of failures might simply be result of a number of combinations we have for k8s tests - in "canary" builds we run I think ~ 30 different combos of the k8s run (they are run for different python versions, kubernetes versions and executors) - this is why a "small" flakiness might multuply by a number of matrix entries we have. One other strategy that might help here is to limit the number of matrix entries we run - we already do it for regular tests - where we do not run the full matrix, but we have some matrix excludes - which results in a somewhat representative subset of full matrix combo - maybe it's a good idea to do something like that **more** here. -- 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]
