Uploaded image for project: 'OpenShift Request For Enhancement'
  1. OpenShift Request For Enhancement
  2. RFE-8717

Allow custom ServiceAccounts or configurable authentication for Cluster Version Operator (CVO) metrics endpoint

XMLWordPrintable

    • Icon: Feature Request Feature Request
    • Resolution: Unresolved
    • Icon: Blocker Blocker
    • None
    • openshift-4.18
    • updates
    • None
    • None
    • Product / Portfolio Work
    • None
    • False
    • Hide

      None

      Show
      None
    • None
    • None
    • None
    • None
    • None
    • None
    • None
    • None

      Proposed title of this feature request:

      1. Allow custom ServiceAccounts or configurable authentication for Cluster Version Operator (CVO) metrics endpoint
      2. What is the nature and description of the request?

      The customer requests a product enhancement to modify the Cluster Version Operator (CVO) metrics endpoint (:9099/metrics) behavior regarding authentication. Currently, this endpoint appears to enforce a strict, hardcoded allow-list that only trusts the specific system identity system:serviceaccount:openshift-monitoring:prometheus-k8s
      The customer requests two specific options to address this:
      1. Disable Authentication: The ability to disable the authentication requirement for the CVO metrics endpoint entirely.
      2. Support Custom ServiceAccounts: The ability for the endpoint to accept custom ServiceAccounts via standard RBAC, removing the limitation of hard-coded identities
      3. Why does the customer need this? (List the business requirements here)
      Third-Party Observability Support: The customer uses Grafana Alloy agents rather than the default in-cluster Prometheus stack to scrape metrics. The current security hardening prevents these custom agents from authenticating, even when assigned the correct ClusterRoles.
      Security and Governance Audit Compliance: The current workaround requires generating and sharing a token for the privileged prometheus-k8s system account with other teams/tools. This creates "audit smell" and security concerns regarding the handling of long-lived system tokens.
      Architectural Flexibility: The customer notes that centralized Prometheus solutions that do not deploy the standard openshift-monitoring stack on every cluster would be unable to scrape these metrics under the current restrictions
      4. List any affected packages or components.
      cluster-version-operator

              rh-ee-smodeel Subin M
              rhn-support-macastil Mario Castillo
              None
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

                Created:
                Updated:
                None
                None