2

I have enabled sentry [1.5.0] with apache airflow [Version 2.2.2] using following configuration in config file.

[sentry]
sentry_on: True
sentry_dsn: [DSN]
sentry_environment: dev

It seems to work just fine for all sort of errors. However Airflow internal errors like Error while health checking kube watcher process. Process died for unknown reasons, broken_pipe,ApiException: too old resource version etc are very frequent (close to 3k errors a day), which eats up my sentry quota very quickly.

While I am not sure how to fix those, I just want to disable those internal errors to be propagated via sentry and only trigger events for DAG execution errors/failures.

Note:The airflow executor I have set is KubernetesExecutor and my airflow is deployed in GKE.

Any help would be appreciated.

JD-V
  • 3,336
  • 1
  • 17
  • 20

0 Answers0