OKD 4.9 - Cluster Autoscaler definition seems not working properly

[gabrilabs75]

Hi all,
on our OKD platform

Client Version: 4.9.0-0.okd-2021-11-28-035710
Server Version: 4.9.0-0.okd-2022-02-12-140851
Kubernetes Version: v1.22.1-1839+b93fd35dd03051-dirty

we have enabled this cluster autoscaler definition

apiVersion: autoscaling.openshift.io/v1
kind: ClusterAutoscaler
metadata:
  creationTimestamp: "2022-01-04T11:39:00Z"
  generation: 2
  name: default
  resourceVersion: "21584331"
  uid: 9def52a4-55c2-4eb4-94c9-b7d3722a9f80
spec:
  balanceSimilarNodeGroups: true
  podPriorityThreshold: -5
  resourceLimits:
    cores:
      max: 431
      min: 8
    maxNodesTotal: 20
    memory:
      max: 766
      min: 4
  scaleDown:
    delayAfterAdd: 10m
    delayAfterDelete: 5m
    delayAfterFailure: 30s
    enabled: true
    unneededTime: 5m

and related machine autoscaler definition

apiVersion: autoscaling.openshift.io/v1beta1
kind: MachineAutoscaler
metadata:
  creationTimestamp: "2022-02-15T17:44:03Z"
  finalizers:
  - machinetarget.autoscaling.openshift.io
  generation: 2
  name: my-project-dev
  namespace: openshift-machine-api
  resourceVersion: "129271702"
  uid: d70f459b-f08a-4cb0-afc8-b723aaaa0947
spec:
  maxReplicas: 5
  minReplicas: 3
  scaleTargetRef:
    apiVersion: machine.openshift.io/v1beta1
    kind: MachineSet
    name: my-project-dev

When a worker node dedicated to this project exceed 95% of cpu and memory becomes Not Ready without scaling up another node as we expect.
As Virtualization Layer we are using RHV 4.4.8.

Any suggestion about this behaviour?

Regards,
Gabriele

[vrutkovs]

Please attach (or upload to the public file sharing service) must-gather archive

[gabrilabs75]

Hi,
we would prefer to avoid sharing sensitive customer cluster data.
can we do it any other way?