Skip to main content

Containerizing Python applications with Docker (Part 2): Flask Dockerfile

Building on **Docker basics**, Flask apps need HTTP servers suited for multiprocess production (gunicorn, waitress, Hypercorn ASGI equivalents) rather than flask run defaults.

📚 Prerequisites​

  • Completed Series 14 (Flask) or equivalent familiarity.

🎯 What you'll learn​

  • Run Flask behind gunicorn inside Docker.
  • Expose configurable workers and ports for orchestrators like Kubernetes.

Example Dockerfile targeting WSGI​

FROM python:3.12-slim

ENV PYTHONDONTWRITEBYTECODE=1 \
    PYTHONUNBUFFERED=1 \
    PORT=8000

WORKDIR /app

COPY requirements.txt .
RUN pip install --no-cache-dir -r requirements.txt

COPY app ./app

RUN pip install gunicorn

CMD ["sh", "-c", "gunicorn -w 4 -b 0.0.0.0:${PORT} app.wsgi:app"]

Assume app/wsgi.py exposes app as the WSGI callable.

Health checks & signals​

Compose/Kubernetes manifests should define health checks hitting /healthz so rolling deploys discard broken pods.

Use docker stop gracefully—ensure base images propagate SIGTERM to worker processes (gunicorn handles this with --graceful-timeout).

💡 Key takeaways​

  • Containers wrap processes, not magically scaled databases—migrate stateful services separately.

➡️ Next steps​

Compare hosted platforms in Cloud deployment (Heroku, AWS, overview).