Using AKS, the cluster has limits that are overcommitted due to kube-system pods having excessively high CPU/memory limits.

  • kube-system pods like coredns or have a combined CPU limit of 300+%
  • This is for a development environment. We have tried several different VM sizes like Standard_B2ms or Standard_D2_v3 which do not seem to affect the kube-system pods limits/configurations.
  • We cannot manually reduce the limits for these kube-system pods because it is subsequently reverted by AKS.
  • Creating an additional node pool did not seem to solve the problem. Node taints were not respected.

kubectl describe node


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.