Details
-
New Feature
-
Resolution: Unresolved
-
Major
-
None
-
None
Description
Our alerts are best practices for production environments, but there will be many cases where there are reasons for the end user to deviate from them. We should have a way of telling CMOS that a failing check is expected and suppress the alert.
Alertmanager already has silences, but a silenced alert will still appear in the dashboards etc (we could possibly just replace all our dashboards with Alertmanager data, though this would need careful consideration). Alternatively, Cluster Monitor has a suppression system that we may be able to take advantage of.