Teardown

2 min read

Learn how to tear down Coder and the infrastructure on which it's deployed.

This guide shows you how to tear down Coder and the cluster it is deployed on.

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 environments and their associated disk space.

Amazon Elastic Kubernetes Service (EKS)

  1. Make sure you're running eksctl version 0.37.0 or later:

    eksctl version
    
  2. List all of the services in your cluster:

    kubectl get svc --all-namespaces
    
  3. Delete any services that have an EXTERNAL-IP value in your namespace:

    kubectl delete svc <service-name>
    
  4. Delete the cluster and its underlying nodes:

    eksctl delete cluster --name <prod>
    

Azure Kubernetes Service (AKS)

  1. Make sure that the environment 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>"
    
  2. Delete the cluster:

    az group delete --resource-group $RESOURCE_GROUP
    

Google Kubernetes Engine (GKE)

  1. Ensure that the environment variables for PROJECT_ID and CLUSTER_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>"
    
  2. Delete the cluster:

    gcloud beta container --project $PROJECT_ID clusters delete \
    $CLUSTER_NAME --zone <zone>
    

Our docs are open source. See something wrong or unclear? Make an edit.