As part of this KEP: https://github.com/kubernetes/enhancements/tree/master/keps/sig-instrumentation/1748-pod-resource-metrics that reached alpha in k8s 1.21 and beta in 1.22, a new endpoint was added to the kube-scheduler to expose more precise pod-level resource requests/limits metrics compared to the container-level ones exposed by kube-state-metrics.

In https://github.com/openshift/cluster-kube-scheduler-operator/pull/283, Clayton added the new target to the scheduler service monitor, so we should consider transitioning to the new metrics in OCP to the precision of our capacity planning dashboards.

KSM metrics:

• kube_pod_container_resource_limits
• kube_pod_container_resource_limits

Scheduler metrics:

• kube_pod_resource_requests
• kube_pod_resource_limits