-
Story
-
Resolution: Unresolved
-
Minor
-
None
-
None
-
False
-
False
-
Undefined
-
Story: As an OpenShift admin I want the Quay Operator to set up reasonable default alerting rules for Quay/Clair deployments so that I get notifications when certain metrics indicate an unhealthy or underperforming Quay registry.
Acceptance criteria:
- resonable default alert rules on hardware resource utilization (CPU, Memory, DiskSpace) for Quay components and dependencies
- default alert rules for Quay overall healthiness (as defined in the Quay health endpoint)
- alerts that make the Quay admin understand that the build workers fight a huge backlog
- alerts that make the Quay admin understand that the mirroring workers fight a huge backlog
- alerts that make the Quay admin understand that clair fight a huge scanning backlog
- alerts on exceeding average / n-th percentile pull / push times
- alerts on exceeding API latencies
- alerts on low available storage space of any of the Quay dependencies (database, K/V store, NooBaa backend)
- relates to
-
PROJQUAY-288 OpenShift Console alerting triggered by various Quay events
- New