Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-22117

HighOverallControlPlaneCPU alert for SNO has wrong threshold

XMLWordPrintable

    • Moderate
    • No
    • 3
    • OCP VE Sprint 244, OCPEDGE Sprint 248, OCPEDGE Sprint 249, OCPEDGE Sprint 250, OCPEDGE Sprint 251, OCPEDGE Sprint 252
    • 6
    • False
    • Hide

      None

      Show
      None
    • After speaking with monitoring team, i'll proceed with measuring cpu usage of control plane components described in the openshift docs

      Description of problem:

      HighOverallControlPlaneCPU alert for SNO is set to the default from MNO, which is 60% of master node capacity. FOR SNO: This leads to either false positive alerts (in small SNO clusters with e.g. only 4 cores), or false negative alerts (e.g no alert on a cluster with 40 cores, and control plane consuming 50% cores).

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

      4.12

      How reproducible:

      always

      Steps to Reproduce:

      1. Install SNO
2. take a look at "$ oc get prometheusrules cpu-utilization -n openshift-kube-apiserver -o yaml | grep HighOverallControlPlaneCPU
..."
 3. see it to be ">60"

      Actual results:

      see it to be ">60"

      Expected results:

      adjusted for SNO to a sensible value

      Additional info:

       

              bzamalut@redhat.com Bulat Zamalutdinov
              dfroehli42rh Daniel Fröhlich
              Pedro Jose Amoedo Martinez Pedro Jose Amoedo Martinez
              Chad Scribner
              Votes:
              0 Vote for this issue
              Watchers:
              17 Start watching this issue

                Created:
                Updated:
                Resolved: