delete-cluster.sh

Previous Next

Purpose

Delete an EKS cluster safely, optionally deleting managed nodegroups first.

Location

cloud/aws/eks/delete-cluster.sh

Preconditions

  • Required tools: bash, aws
  • Required permissions: eks:DeleteCluster, optional eks:ListNodegroups, eks:DeleteNodegroup, eks:Describe*
  • Required environment variables: none

Arguments

Flag Required Default Description
--name NAME Yes N/A Cluster name
--delete-nodegroups No false Delete all nodegroups before cluster delete
--if-missing No false Return success if cluster absent
--wait / --no-wait No wait enabled Wait for deletion
--timeout SEC No 1800 Wait timeout
--poll-interval SEC No 20 Wait interval
--region REGION, --profile PROFILE No AWS defaults AWS context override
--yes Cond. false Required for non-dry-run deletion
--dry-run No false Print planned commands

Scenarios

  • Happy path: nodegroups (optional) and cluster are deleted cleanly.
  • Common operational path: ephemeral cluster teardown in integration environments.
  • Failure path: deletion blocked by remaining resources or missing permissions.
  • Recovery/rollback path: remove blockers manually, rerun deletion sequence.

Usage

cloud/aws/eks/delete-cluster.sh --name dev-eks --yes
cloud/aws/eks/delete-cluster.sh --name dev-eks --delete-nodegroups --yes
cloud/aws/eks/delete-cluster.sh --name dev-eks --if-missing --dry-run

Behavior

  • Main execution flow:
  • validates cluster existence (or skip via --if-missing)
  • optionally deletes nodegroups first
  • submits cluster deletion request
  • optionally waits until cluster no longer exists
  • Idempotency notes: convergent with --if-missing.
  • Side effects: destructive removal of cluster control-plane resources.

Output

  • Standard output format: timestamped deletion logs.
  • Exit codes:
  • 0 success
  • 2 invalid input or missing --yes
  • non-zero AWS API failures/timeouts

Failure Modes

  • Common errors and likely causes:
  • missing --yes safety confirmation
  • nodegroups not deleted when required by cluster state
  • IAM denies delete actions
  • Recovery and rollback steps:
  • rerun with --delete-nodegroups
  • verify IAM permissions and retry
  • inspect EKS events and residual dependencies

Security Notes

  • Secret handling: none.
  • Least-privilege requirements: destructive EKS permissions should be restricted to teardown roles.
  • Audit/logging expectations: cluster teardown operations must be change-approved and auditable.

Testing

  • Unit tests:
  • safety guard and wait mode logic
  • Integration tests:
  • teardown in disposable test clusters
  • Manual verification:
  • aws eks describe-cluster returns not found post-delete
create-cluster.sh upgrade-cluster.sh