EKS Cluster Deletion Process

EKS Cluster Deletion Process A workflow diagram generated by Archify. 01 / Pre-cleanup 02 / Cluster deletion Remove LBs · Service · Ingress · Pre-cleanup Remove LBs Service · Ingress Delete PVCs · Reclaim volumes · Pre-cleanup Delete PVCs Reclaim volumes Delete node groups · EC2 terminated · Pre-cleanup Delete node groups EC2 terminated Fargate profile · Delete profile · Pre-cleanup Fargate profile Delete profile Delete cluster · delete-cluster · Cluster deletion Delete cluster delete-cluster Leftovers · VPC · IAM · Cluster deletion Leftovers VPC · IAM 1 storage 2 nodes 3 serverless 4 cluster 5 verify Legend Workload cleanup Final deletion AWS resources

Why the order matters

  • • Leave LoadBalancer Services and Ingresses behind and the ALBs and NLBs keep billing
  • • Deleting PVCs first lets the reclaim policy clean up the EBS volumes
  • • Cluster deletion fails while node groups or Fargate profiles still exist

Commands

  • • kubectl delete svc --all-namespaces --field-selector spec.type=LoadBalancer
  • • eksctl delete cluster --name my-cluster removes stacks in the right order
  • • If you built it with the CLI, delete in this order: delete-nodegroup → delete-fargate-profile → delete-cluster

After deletion

  • • A hand-built VPC, security groups, and IAM roles survive and need separate cleanup
  • • Check CloudWatch log groups and EBS snapshots against your retention policy
  • • Back up anything you need before deleting and confirm it can be restored