Closed hjacobs closed 8 years ago
So "generally" the single entity removal cleanup works. but due to race conditions (e.g. AWS scheduler/worker hasn't pushed results yet), it might not fail sometimes.
We should implement a "brute force" cleanup to eventually remove all stale alerts/check results.
https://github.com/zalando-zmon/zmon-scheduler/tree/feature-cleanup-alert-change
Finally doing the cleanup after alert definition changes
Apparently only the old Python scheduler still does this...