Execute o Flask com Gunicorn

Gunicorn is a production WSGI server for Python. It runs your Flask app across multiple worker processes so many users can be served at once. 🦄

Add Gunicorn to your environment with pip. Treat it as a real dependency and pin it in your requirements file.

Gunicorn must find the WSGI callable. That is the Flask app instance you created, usually named app in your module.

You point Gunicorn at your app using module:variable syntax. The part before the colon is the file, the part after is the app name.

Many teams add a tiny wsgi.py that just exposes the app. It keeps the run command stable even as the app grows.

The --workers flag sets how many processes handle requests. A common starting point is two times your CPU cores plus one.

Use --bind to choose the host and port. Binding to 0.0.0.0 lets the app accept traffic from outside the container.

Using create_app? Gunicorn can call your factory directly with a function call in the target string.

The --timeout flag kills a worker stuck on a slow request. It stops one bad request from freezing the whole server.

Gunicorn does not auto-reload on code changes in production. You restart it on deploy, which gives you a predictable release moment.

Gunicorn writes structured access logs showing each request and its status. Pipe them to stdout so your platform can collect them.

How do you tell Gunicorn which app to run?

Gunicorn runs your Flask app with real workers, a bind address, and timeouts. Next you will lock all of this into a Docker image. 📦