StackLight

StackLight

  • [35572] Fixed the issue with Alerta updating incorrectly due to invalid API keys, causing alerts to get stuck.

  • [35375] Fixed the memory leak issue causing Prometheus Relay to allocate all available RAM on the node.

  • [35533] Fixed the issue with the Nova Compute graph in the Nova - Overview Grafana dashboard displaying incorrect information.

  • [35683] Fixed the issue with Telegraf failing to gather the compute quota set metrics.

  • [35549] Fixed the issue with improper monitoring of public endpoint certificates. Replaced the unavailability alerts for the OpenStack components endpoints with a generic OpenstackServiceEndpointDown alert. Enhanced the OpenStack Grafana dashboards to display the public endpoint.

  • [35536] Improved the GlusterFs monitoring by adding the GlusterfsMountMissing that raises when a GlusterFS mount point is not mounted.

  • [35435] To avoid issues with short-living alerts, added the for: 2m clause to the alerts that use the procstat_running metric.

  • [35685] Fixed the raise condition of the HaproxyBackendDownMajor alert to properly indicate the number of back ends in the DOWN state.

  • [35477] To avoid issues with Keystone authorization errors due to keys being out of sync, added the KeystoneKeysRotationFailure alert to monitor the rotation of fernet and credentials Keystone keys.