AAS

A webhook listener for use at Study Association Sticky, written as a Flask application.

Its current use is to make automatic deployment to our server possible of our website and our sign-up page, triggered by GitHub pushes or changes in the Contentful CMS it uses.

This project can be extended for other webhook processing.

Usage

This project requires uv to install the right python version and packages in a virtual environment.

Development

# Use the provided 'sample.config.json' to create a 'config.json'
# file and adjust it to your needs. (See 'Configuration' section)
cp sample.config.json config.json
vim config.json
# Run the development server using uv
uv run aas.py

Configuration

Aas is build to support multiple types of webhooks. For every type of endpoint you can define a class (webhook handler) which will handle all requests to that endpoint. For example, you can define a webhook handler which runs a systemd service on incoming webhooks.

To abstract your endpoints from this implementation, aas loads up your endpoint config from config.json. This json contains a mappings from endpoints to webhook handlers. This means that if you ever want to add an endpoint, you simply expand the config.json.

Production

The dependencies include gunicorn, which is a WSGI server for use in production. You can use the following commands as a simple example:

uv run gunicorn aas:aas

This binds gunicorn to http://localhost:8000/. You should place a reverse proxy, like nginx, in front of this.

To set this up in our own production environment, we use some Ansible tasks.

Testing consumption of webhooks in development

To make the development server temporarily available for webhook consumers in the outside world (e.g. GitHub), you can use ngrok.

To use it, you can run aas in one terminal and in another terminal run ngrok http 5000 (5000 being the default port Flask's development server binds to).