Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-50004

Correlated query returns empty result due to missing cluster:master_nodes metric

XMLWordPrintable

    • +
    • Low
    • None
    • False
    • Hide

      None

      Show
      None
    • Release Note Not Required
    • In Progress

      Description of problem:

      After following the procedure described at https://github.com/openshift/installer/blob/main/docs/user/openstack/observability.md the PromQL query:

      sum by (vm_instance) (
  group by (vm_instance, resource) (ceilometer_cpu)
    / on (resource) group_right(vm_instance) (
      group by (node, resource) (
        label_replace(kube_node_info, "resource", "$1", "system_uuid", "(.+)")
      )
    / on (node) group_left group by (node) (
      cluster:master_nodes
    )
  )
)

      returns an empty result because the cluster:master_nodes metric is not being scraped from the shift-on-stack cluster.

      Version-Release number of selected component (if applicable):

      4.18.0-rc.6

      How reproducible:

      Always

      Steps to Reproduce: https://github.com/openshift/installer/blob/main/docs/user/openstack/observability.md

      Actual results:

      Empty result

      Expected results:

      The number of OpenShift master nodes per OpenStack host

      Additional info:

      Just querying the 'cluster:master_nodes' metric returns empty

      The procedure described in the docs needs to be updated and include the cluster:master_nodes metric retrieval in the scrapeconfig definition.

      spec:
  params:
    'match[]':
    - '{__name__=~"kube_node_info|kube_persistentvolume_info|cluster:master_nodes"}'

              juriarte@redhat.com Jon Uriarte
              juriarte@redhat.com Jon Uriarte
              Itshak Brown Itshak Brown
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Created:
                Updated: