You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently (without non-disruptive vertical pod autoscaling in use) this is a symptom of the node's kubelet running with CPUManager enabled and configured with a policy other than none (IOW, static). The CPUManager has a reconcilation loop which is periodically sending update requests for all containers running on the node. CRI Resource Manager ignores these.
Please check your cluster's/kubelet's configuration and set the built-in CPU policy to none.
Describe the bug
There are a lot of messages in the cri logs stating:
Does this mean the current configuration is not yet implemented or can these warnings be ignored?
Expected behavior
No warnings.
To Reproduce
Environment
Additional context
The text was updated successfully, but these errors were encountered: