Skip to content

Correct walgBackupDelayed alert for patroni-embedding

Devin Sylva requested to merge devin/walg_timing_for_embedding into master

To correct a frequent alert which is caused by low traffic on the embedding database, we need to allow more time between walg backups on those clusters. https://gitlab.com/gitlab-com/gl-infra/reliability/-/issues/23831

It seems like the easiest approach is to split the expression to not check the embedding database for the 15 minute timing, and add an or statement checking only the embedding database and allowing 60 minutes (which we can then easily adjust later if it's still too short)

Edited by Devin Sylva

Merge request reports