4.18.0-okd-scos.0: unexpected HighOverallControlPlaneCPU and ExtremelyHighIndividualControlPlaneCPU alerts

[soubinan]

Hello,
I notice the HighOverallControlPlaneCPU and ExtremelyHighIndividualControlPlaneCPU alerts on my sno since I upgraded it from 4.17.0-okd-scos.1

Infrastructure type: SNO on barre metal
CPU: 112
Memory: 256GB

I noticed the issue is coming from the ovnkube-controller container (a pod from ovnkube-node)

Nothing evident to point to in the container's logs. It seems working as expected (actually I don't know) but I have the feeling it is over processing something when seeing the logs.

Note the platform has no custom workload for now since it is new. I just added some operators since the initial installation (cert manager, virtualization, lvms, odf)

I am not sure if it is expected considering I have no real traffic on this sno, and I am worried about how it would behave once it has more to process

EDIT:
Adding the logs

onvkube-controller.log
kube-apiserver.log

[HAMA2007]

I have a same problem.

ovnkube-node is using x10~100 MORE CPU than old OKD.

[ Left OKD 4.18+SCOS / RIGHT OKD 4.15+FCOS ]

[soubinan]

I was not able to find what the issue root cause was
But all goes to the normal after I upgraded my node from 4.18.0-okd-scos.0 to 4.18.0-okd-scos.1