=====================================
Local testing of the Kubernetes stack
=====================================

These are some brief but essential steps to get everything from cloud-api to
cloud-k8s-operator running locally (each of them requires a Python virtualenv
with dependencies installed).

- Start the Brain and Redis:

  .. code-block:: console

     $ cd path/to/your/projects/crate/cloud
     $ docker-compose up -d brain redis

- Run cloud-api

  .. code-block:: console

     $ cd path/to/your/projects/crate/cloud-api
     $ source path/to/cloud-api-venv/bin/activate
     (cloud-api) $ FLASK_LOGGING=DEBUG FLASK_HOST=127.0.0.1 FLASK_PORT=8080 ENV=local cloud-api

- Run cloud-agent

  .. code-block:: console

     $ cd path/to/your/projects/crate/cloud-agent
     $ source path/to/cloud-agent-venv/bin/activate
     (cloud-agent) $ cloud-agent

- Run cloud-api-celery

  .. code-block:: console

     $ cd path/to/your/projects/crate/cloud-api
     $ source path/to/cloud-api-venv/bin/activate
     (cloud-api) $ CELERY_LOG_LEVEL=DEBUG LOCAL_ENV=dev/local.env celery worker --app cloud.api.celery.worker.celery --concurrency 1

  Depending on the IP address you'll run cloud-agent on, you might need to
  change the value for `CLOUD_API_AGENT_URL` in `dev/local.env`.

- Run Minikube and make sure that's your default Kubernetes context

  .. code-block:: console

     $ minikube start
     $ kubectl config use-context minikube

- Run the Kubernetes operator for projects

  .. code-block:: console

     $ cd path/to/your/projects/cloud-k8s-operator
     $ source path/to/cloud-k8s-operator-venv/bin/activate
     (cloud-k8s-operator) $ kopf run --verbose --standalone ./cloud/operator/project.py