The kube-apiserver pods are consuming high memory and CPU in RHOCP 4
Issue
- Observing
kube-apiserver
process consumes noticeable gigabytes of memory. - Observing a high load on the cluster.
- Pods are taking a long time to start.
- The control plane nodes are getting stuck and becoming
NotReady
. - Experiencing a sudden memory usage increase on
kube-apiserver
pod causing control plane node to be unresponsive. - The etcd become not healthy after control plane node memory usage reached high values.
Environment
- Red Hat OpenShift Container Platform (RHOCP)
- 4
kube-apiserver
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.