This guide shows you how to tear down Coder and the cluster on which it is deployed.
These instructions help you remove infrastructure created when following our Kubernetes setup tutorials. They do not include teardown steps for any additional resources that you create. If you need to keep your cluster, you can run
helm uninstall coder
, which deletes all Coder services but retains workspaces and their associated disk space.
Amazon Elastic Kubernetes Service (EKS)
-
Make sure you're running
eksctl
version 0.37.0 or later:eksctl version
-
List all of the services in your cluster:
kubectl get svc --all-namespaces
-
Delete any services that have an
EXTERNAL-IP
value in your namespace:kubectl delete svc <service-name>
-
Delete the cluster and its underlying nodes:
eksctl delete cluster --name <prod>
Azure Kubernetes Service (AKS)
-
Make sure that the workspace variable for
RESOURCE_GROUP
is set to the one you want to delete in Azure:echo $RESOURCE_GROUP
If the variable is incorrect, fix it by setting it to the proper value:
RESOURCE_GROUP="<MY_RESOURCE_GROUP_NAME>"
-
Delete the cluster:
az group delete --resource-group $RESOURCE_GROUP
Google Kubernetes Engine (GKE)
-
Ensure that the environment variables for
PROJECT_ID
andCLUSTER_NAME
are set to those for the cluster you want to delete:echo $PROJECT_ID echo $CLUSTER_NAME
If these values are incorrect, you can fix this by providing the proper names:
PROJECT_ID="<MY_RESOURCE_GROUP_NAME>" \ CLUSTER_NAME="<MY_CLUSTER_NAME>"
-
Delete the cluster:
gcloud beta container --project $PROJECT_ID clusters delete \ $CLUSTER_NAME --zone <zone>