Add alert rules for missing metrics

Metrics can be missing due to cache expiration or connectivity issue
between the exporter and OpenStack APIs. Also see issue [1]

[1]: https://github.com/canonical/openstack-exporter-operator/issues/130

Change-Id: I081b7d6f704953ef3360ab6bbed20dcb2c59ec9f
Signed-off-by: Chi Wai Chan <chiwai.chan@canonical.com>
This commit is contained in:
Chi Wai Chan 2025-03-10 17:25:25 +08:00
parent 8cf7d21841
commit e632c03547
No known key found for this signature in database

View File

@ -14,3 +14,15 @@ groups:
summary: OpenStack Services Down
description: |
The OpenStack service {{ $labels.service }} is down
- name: OpenStackMetrics
rules:
- alert: OpenStackMetricsMissing
expr: |
absent_over_time({__name__=~"openstack_(.+)_up"}[5m])
labels:
severity: critical
annotations:
summary: OpenStack Metrics Missing
description: |
All OpenStack metrics are missing for over 5 minutes. This could be due to the
connectivity issue of the OpenStack APIs, or the cache of the metrics has expired.