fardjad/webhook-deploy repository overview

⁠Webhook Deploy

⁠Synopsis

An Alpine Linux⁠ based image that can clone a Git repository and execute a custom setup script upon receiving webhooks.

In order to keep the image small and as close to the base image as possible, no system-wide package is installed; Git and SSH clients are compiled, statically linked, and placed in their own directories.

⁠Environment Variables

The following environment variables must be set before starting the container

1. SSH_PRIVATE_KEY: a string containing the private key (in OpenSSH format) to access the Git repository
2. GIT_REPO_URL: Git repository URL
3. GIT_REPO_BRANCH: Git branch to checkout before running the setup script

In order to prevent unauthorized clients from triggering the webhook handler, one can set WEBHOOK_AUTH_TOKEN environment variable. The webhook receiver server will only accept URL paths that end in WEBHOOK_AUTH_TOKEN.

⁠How it Works

Webhook receiver server listens on port 8000 and does the following upon receiving webhooks:

1. Clones the specified Git repository
2. Checks out the specified branch
3. Looks for .webhook/setup script within the cloned repository and executes it

⁠Notes

1. Only SSH protocol is supported for cloning Git repositories
2. .webhook/setup script must have a shebang line (ex. #!/bin/sh)
3. Webhook receiver server will clone the repository to a temporary directory and removes it after setup script is finished
4. Setup script will be executed relative to the repository directory
5. Nothing should be executed directly from the setup script's working directory
6. It's up to the setup script to compile/process and copy required files to their correct location before the cloned repository directory gets removed
7. It's up to the setup script to start and track the state of services within the container
8. Setup script should not start any blocking foreground processes
9. It's highly recommended to use nohup to start background processes in setup script

⁠License

MIT⁠