Uploaded image for project: 'Red Hat Advanced Cluster Management'
  1. Red Hat Advanced Cluster Management
  2. ACM-27834

Observability tries to determine if a cluster is OCP too early in deployment

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Unresolved
    • Icon: Undefined Undefined
    • None
    • ACM 2.14.1, MCE 2.9.1
    • Observability
    • None
    • Important
    • Customer Escalated
    • None

      Description of problem:

      Observability relies on the openshiftVersion tag to determine if a cluster is OCP or not but this tag is only set after the clusterversion event

          history:
          - completionTime: null
            image: quay.io/openshift-release-dev/ocp-release@sha256:df00d64554b9b5d860273f938e0ff3b0c4a7af8356fbd103658ebe6fa2a830c8
            startedTime: "2025-12-10T14:35:36Z"
            state: Partial
            verified: false
            version: 4.17.29
          observedGeneration: 1
          versionHash: YF5MJ2ozP5A=
      

      Because observability is getting deployed on the cluster as a non-ocp, which is causing issues with the deployment itself, so the label will not get applied.

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

      ACM 2.14.1
      MCE 2.9.1
      OCP 4.17.29

      How reproducible:

      all the time

      Steps to Reproduce:

      1. enable observability
      2. deploy hcp on a child MCE
      3. ...

      Actual results:

      The deployment is stuck with monitoring failing due to alertmanager-main-0 failing to start ; this in turn is due to observability trying to manage alertmanager on a ocp cluster

      Expected results:

      Observability only tries to deploy after the clusterversion state is "completed":

        history:
        - completionTime: "2025-12-19T05:39:03Z"
          image: quay.io/openshift-release-dev/ocp-release@sha256:df00d64554b9b5d860273f938e0ff3b0c4a7af8356fbd103658ebe6fa2a830c8
          startedTime: "2025-12-19T04:47:48Z"
          state: Completed
          verified: false
          version: 4.17.29
        observedGeneration: 1
        versionHash: YF5MJ2ozP5A=
      

      Additional info:

      During the deployment, observability tries to handle the cluster as a non-OCP cluster :
      oc set env open-cluster-management-addon-observability/endpoint-observability-operator --list shows that INSTALL_PROM=True

      related to jira ACM-27803

      hundreds of clusters are being deployed in the enviroment - it may be needed to deploy several at the same time to reproduce the issue.

      workaround : https://access.redhat.com/solutions/7135219

              Unassigned Unassigned
              rhn-support-fdewaley Felix Dewaleyne
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated: