Dockerize Django: Gunicorn, Celery, and Static Files Done Right

Django deployments require coordinating several moving parts: the WSGI server, Celery workers, a message broker, static file collection, and database migrations. Without containers, this means Supervisor configs, virtualenvs that may or may not match, and a deployment script that runs collectstatic and migrate - hopefully in the right order.

With Docker, each component runs in its own container from the same image. Migrations run as an init container or entrypoint step. Static files are collected at build time. The entire stack spins up with docker compose up.

Multi-stage Dockerfile with Compose orchestration:

• Build stage: FROM python:3.12-slim AS builder - installs build deps, creates virtualenv, runs pip install -r requirements.txt, then python manage.py collectstatic --noinput.
• Runtime stage: FROM python:3.12-slim - copies virtualenv and collected static files. Installs only runtime system deps.
• Compose services: web (gunicorn myapp.wsgi:application --bind 0.0.0.0:8000 --workers 4), celery-worker (celery -A myapp worker -l info), celery-beat (celery -A myapp beat -l info), redis, postgres.

The entrypoint runs python manage.py migrate --noinput before starting Gunicorn. Health check hits /health/ with Django's health check middleware. Static files are served via whitenoise (zero-config) or an Nginx sidecar for high-traffic apps.

• Multi-stage Dockerfile with virtualenv isolation - 150-250 MB final image
• docker-compose.yml with web, Celery worker, Celery beat, Redis, and Postgres
• Entrypoint script running migrations and collectstatic
• Gunicorn config with worker count, timeout, and graceful shutdown
• Whitenoise setup for static files (or Nginx config for heavy traffic)
• CI pipeline with pip cache and Docker layer optimization