Dockerize a Python Flask Web Application

Python remains one of the most widely deployed languages for web APIs, and Flask is its most approachable framework for building small services. But moving a Flask app from python app.py on a laptop to a portable, reproducible container is where most developers hit friction: pip layers rebuild on every code change, the dev server hangs on shutdown, __pycache__/ directories leak into bind mounts, and flask run ends up in production because nobody explained why gunicorn matters. A well-authored Dockerfile fixes all of this with a handful of well-ordered instructions.

This lab walks through the same Dockerfile workflow real Python teams use in production. You will start with a pre-built Flask API in your workspace, write a naive Dockerfile that works but rebuilds slowly, then refactor it to leverage layer caching, add a .dockerignore, split the image into a dev variant with hot reload and a production variant that uses gunicorn as the WSGI server with a non-root user. By the end you will understand exactly why each line of a Python Dockerfile is shaped the way it is, and what changes between development and production.

Objectives

Upon completion of this beginner level lab, you will be able to:

• Author a naive Dockerfile that copies a Flask app, installs requirements.txt, and runs the app on a non-reserved port.
• Refactor the Dockerfile to copy requirements.txt before the application code so pip layers stay cached across code edits.
• Write a .dockerignore that excludes __pycache__/, *.pyc, .venv/, .git/, and .env from the build context.
• Build a Dockerfile.dev that enables Flask debug mode and hot reload via a bind mount, with PYTHONDONTWRITEBYTECODE=1 to prevent .pyc leakage to the host.
• Build a Dockerfile.prod that runs gunicorn as a non-root user using exec-form CMD for clean SIGTERM handling.

Who is this lab for?

This lab is designed for:

• Developers with basic Python and Flask familiarity who want to package a Python web service for production.
• Engineers who have written Dockerfiles in other languages and want to learn the Python-specific patterns around pip caching, bytecode, and WSGI servers.
• Platform learners moving from python app.py on a laptop to a portable, reproducible image.

No prior Docker knowledge is required, but learners should understand pip, virtual environments, and the difference between a development server and a production server.