Description of problem: The documentation for the metric job:kube_pv:count is wrong. It only counts ODF PVs. The metric is created by ODF here: https://github.com/red-hat-storage/ocs-operator/blob/main/controllers/storagecluster/prometheus/localcephrules.yaml#L24-L26 I've already opened a Github issue about this a while ago... it has become rotten by now https://github.com/openshift/cluster-monitoring-operator/issues/1586 Version-Release number of selected component (if applicable): OCP & ODF 4 (has been that way for a while now) How reproducible: Always when reading metric documentation at https://github.com/openshift/cluster-monitoring-operator
Thanks for filing a BZ. The documentation is certainly incorrect, I would argue so is the metric/recording rule name. A better name would indicate that the PVs are ceph based, e.g. job:kube_ceph_pv:count. Assigning to the ODF team as they own this rule.
issue is fixed in 4.12, see: https://github.com/openshift/cluster-monitoring-operator/blob/release-4.12/manifests/0000_50_cluster-monitoring-operator_04-config.yaml#L343-L352 "job:kube_pv:count" is deprecated in ODF 4.12, does that also mean we could remove it since OCP 4.13?
I would vote to remove the old metric