0

so I'm using eks amazon managed cluster for Kubernetes , I noticed some of the pods are using too much resources ,memory and cpu.

pods are kube-controller and kube-flannels

enter image description here

current version of Kubernetes is 1.22 (did upgrade from 1.21 bc of the issue)

my question here is that normal, if not what're you recommendations to solve this issue.

Regards,

  • What is the pattern for this? Is it happening over a long period of time? is the usage increasing with time? staying the same? This simply means your pods are making a lot of communication with eachother which is normal. It becomes a problem when you start running out of resources and new pods are not scheduled on nodes, but that doesn't seem to be the case here. – zer0 Jul 26 '22 at 21:48
  • Base on the screen shot this cluster is not EKS. – gohm'c Jul 27 '22 at 01:12
  • it's actually the same , consistent over time , once I restarted the pods( kube systems pods) , it will increase drastically until it reaches the limit I'm showing in the image – Turki Alhazzani Jul 27 '22 at 09:00
  • gohm'c it's eks man – Turki Alhazzani Jul 27 '22 at 09:00

1 Answers1

0

Answering my question , I have no idea what was happening , I just deleted the cluster and created new one with version 1.22 , and everything seems working fine.