Production ASGI Architecture: Gunicorn with Uvicorn Workers

1Concept

Gunicorn acts as the master process manager (handling worker restarts, OS signals, and socket bindings), while Uvicorn worker classes (`UvicornWorker`) execute the asynchronous event loops.

2Architecture Diagram

Gunicorn Master Process (Monitors workers, binds port 8000)
  ├── Worker 1 (Uvicorn Worker - Async Event Loop)
  ├── Worker 2 (Uvicorn Worker - Async Event Loop)
  └── Worker 3 (Uvicorn Worker - Async Event Loop)

3Code Example

gunicorn_cmd = '''
gunicorn main:app \
    --workers 4 \
    --worker-class uvicorn.workers.UvicornWorker \
    --bind 0.0.0.0:8000 \
    --timeout 60 \
    --access-logfile - \
    --error-logfile -
'''
print("=== Production Gunicorn + Uvicorn Deployment ===")
print(gunicorn_cmd.strip())

4Expected Output

=== Production Gunicorn + Uvicorn Deployment ===
gunicorn main:app \
    --workers 4 \
    --worker-class uvicorn.workers.UvicornWorker \
    --bind 0.0.0.0:8000 \
    --timeout 60 \
    --access-logfile - \
    --error-logfile -

5Key Takeaways

• ✓Rule of thumb for workers: `workers = (2 * CPU_cores) + 1`.
• ✓Gunicorn gracefully restarts workers that exceed memory limits (`--max-requests 1000`).
• ✓Logs are streamed directly to stdout (`-`) for container aggregation (Fluentd/Datadog).