Reads the Kubernetes clusters in one location (optionally filtering by labels) and clones them into another (or just outputs JSON as a dry run), to/from AWS, GCP, and Azure.

The Cluster Cloner project is discussed in this article, and the different cluster models are compared.

For usage, run clustercloner --help

• Add a file credentials-cluster-manager.json (or another file name) with credentials for a service account with the Kubernetes Cluster Admin role (to read and create clusters).
  • This is loaded through the GOOGLE_APPLICATION_CREDENTIALS environment variable, which you should set to credentials-cluster-manager.json (or other file name). If this environment variable is not set, your logged-in gcloud account will be used.
• Add a file .env with Azure credentials.
  • Use .env.tpl as a template.
  • The user should have the Azure Kubernetes Service Cluster Admin Role.
• Add a file awscredentials with AWS credentials.
  • The application uses the AWS_SHARED_CREDENTIALS_FILE environment variable to find this file, but awscredentials is the default.
  • Use awscredentials.tpl as a template.
  • The user should have the policy discussed here. Specific example here

• Not needed for local build.
• Store the base-64 encodings, for example echo my-credential.json |base64
• For GitHub CI, please specify the following Secrets in GitHub:
  • AZ_ENV_BASE64 .env file with Azure credentials for the application, following .env.tpl as a template.
  • AWS_CREDENTIALS_BASE64 AWS credentials file for the application. Use awscredentials.tpl as a template.
  • GCP_CLUSTER_MANAGER_KEYJSON_BASE64 Google credentials file (JSON) for the application.
  • GCR_PUSHER_KEYJSON_BASE64 Google credentials file (JSON) with role Storage Admin for pushing to your GCR registry
  • DOCKER_REGISTRY - Registry to which the GitHub action with authenticate, for example gcr.io.
  • DOCKER_REPOSITORY - optional. For example gcr.io/my-gcp-project/image-name. If missing, the value defaults to $GITHUB_REPOSITORY, in the form user/repo.

The Docker image is built in Github Workflows. In development, you can run DOCKER_BUILDKIT=1 docker build -t <TAG> . (Don't forget that final dot.)

In the GitHub actions, make is used to build the application.

In development:

Plain make to format, lint and build

make lint and make fmt

make test-short

make test. Before running, make sure that your Google Project and Azure Group are set up (see "Credentials" above) and have the same name. The full suite can take up to an hour.

This project was started from the goapp template, a bootstrap project for Go CLI applications.