Project Syn: Commodore

This repository is part of Project Syn. For documentation on Project Syn and this component, see https://syn.tools.

See GitHub Releases for changelogs of each release version of Commodore.

See DockerHub for pre-built Docker images of Commodore.

Commodore is published on PyPI

Overview

Commodore provides opinionated tenant-aware management of Kapitan inventories and templates. Commodore uses Kapitan for the heavy lifting of rendering templates and resolving a hierachical configuration structure.

Commodore introduces the concept of a component, which is a bundle of Kapitan templates and associated Kapitan classes which describe how to render the templates. Commodore fetches any components that are required for a given configuration before running Kapitan, and sets up symlinks so Kapitan can find the component classes.

Commodore also supports additional processing on the output of Kapitan, such as patching in the desired namespace for a Helm chart which has been rendered using helm template.

System Requirements

• Python 3.8 - 3.11 with python3-dev and python3-venv updated
• jsonnet-bundler
  • Our fork projectsyn/jsonnet-bundler is currently recommended. It parallelizes fetching of dependencies, which speeds up Commodore significantly, and has fixes to make the dependency fetching more deterministic.
• libmagic (install with brew install libmagic on macOS)

Getting started

1. Recommended: create a new virtual environment

   python3 -m venv venv
   source venv/bin/activate

2. Install commodore from PyPI

   pip install syn-commodore

3. Download jsonnet-bundler from projectsyn/jsonnet-bundler/releases and put the binary in your $PATH as jb.

4. For Commodore to work, you need to run an instance of Lieutenant somewhere (locally is fine too).

5. Setup a .env file to configure Commodore (don't use quotes):

   # URL of Lieutenant API
   COMMODORE_API_URL=https://lieutenant-api.example.com/
   # Lieutenant API token
   COMMODORE_API_TOKEN=<my-token>
   # Your local user ID to be used in the container (optional, defaults to root)
   USER_ID=<your-user-id>
   # Your username to be used in the commits (optional, defaults to your local git config)
   COMMODORE_USERNAME=<your name>
   # Your user email to be used in the commits (optional, defaults to your local git config)
   COMMODORE_USERMAIL=<your email>

6. Run commodore

   commodore

Run Commodore with poetry

Additional System Requirements

• Poetry 1.3.0+
• Docker

1. Install requirements

   Install poetry according to the upstream documentation.

   Create the Commodore environment:

   poetry install

   Download jsonnet-bundler from projectsyn/jsonnet-bundler/releases and put the binary in your $PATH as jb.

2. Finish setup as described above

3. Run Commodore

   poetry run commodore

4. Start hacking on Commodore

   poetry shell

   • Write a line of test code, make the test fail
   • Write a line of application code, make the test pass
   • Repeat

   Note: Commodore uses the Black code formatter, and its formatting is encforced by CI.

5. Run linting and tests

   Auto format with autopep8

   poetry run autopep

   List all Tox targets

   poetry run tox -lv

   Run all linting and tests

   poetry run tox

   Run just a specific target

   poetry run tox -e py38

Run Commodore in Docker

IMPORTANT: After checking out this project, run mkdir -p catalog inventory dependencies in it before running any Docker commands. This will ensure the folders are writable by the current user in the context of the Docker container.

A docker-compose setup enables running Commodore in a container. The environment variables are picked up from the local .env file. By default your ~/.ssh/ directory is mounted into the container and an ssh-agent is started. You can skip starting an agent by setting the SSH_AUTH_SOCK env variable and mounting the socket into the container.

1. Build the Docker image inside of the cloned Commodore repository:

docker-compose build

1. Run the built image:

docker-compose run commodore catalog compile $CLUSTER_ID

Documentation

Documentation for this component is written using Asciidoc and Antora. It is located in the docs/ folder. The Divio documentation structure is used to organize its content.

Run the make docs-serve command in the root of the project, and then browse to http://localhost:2020 to see a preview of the current state of the documentation.

After writing the documentation, please use the make docs-vale command and correct any warnings raised by the tool.

Contributing and license

This library is licensed under BSD-3-Clause. For information about how to contribute see CONTRIBUTING.