Check manual page of kube_cpu
|Distribution||Official part of Checkmk|
Requests utilization is the CPU usage divided by the sum of all configured container CPU requests. Limits utilization is computed in the same manner. The utilization percentage values are omitted if any of the associated request/limit value are not configured. Moreover, limits utilization is omitted if any container has a limit of zero. The reason for omission is displayed in the check's details.
Some care must be taken, when interpreting the usage on individual nodes. The "CPU resources" service reports the usage of running pods only. This means that the usage of operating and Kubernetes system daemons is not reflected. If you need full usage, use the "CPU utilization" service instead. Another caveat is that "Node utilization" and "Cluster utilization" are computed relative to the allocatable number of cores of the nodes, not relative to their capacity. For details on this distinction please consult the Kubernetes docs.
The service includes metrics for the CPU usage, the requests utilization, the limits utilization, the CPU requests and CPU limits of the Kubernetes object.
There are no default thresholds for this service. However, thresholds can be configured via the "Kubernetes CPU resource utilization" ruleset.
The check requires that the special agent kube is configured.