In my Kubernetes deployment of 3 replicas, I'm doing load testing. I notice that the 3 replicas/pods are not evenly balanced. One pod is regularly maxing out on its CPU limit and other pods are really low usage. I'm using K8s version 1.18.
Any tips and advice on how (if possible) to make it balanced? Note: I am not concerned with any external load balancer nor any nginx controller. This is an internal K8s service call.