Pods failing with OutOfcpu or OutOfmemory in OpenShift 4

Issue

• There are pods failing with reason OutOfcpu or OutOfmemory, even if they have configured requests and limits.
• Some pods are failing to run after they are scheduled in a node.

• There are pods showing OutOfcpu or OutOfmemory messages:

  application-pod           0/1     OutOfcpu     0          2h
  application2-pod          0/1     OutOfmemory  0          2h
  

    message: 'Pod was rejected: Node didn't have enough resource: cpu, requested: 200, used: 12769, capacity: 12800'
    phase: Failed
    reason: OutOfcpu
  

  2h        Normal    Scheduled                      pod/application-pod                         Successfully assigned my-project/application-pod to [node_name]
  2h        Warning   OutOfcpu                       pod/application-pod                         Node didn't have enough resource: cpu, requested: 200, used: 12769, capacity: 12800