Uvicorn and Gunicorn for FastAPI

Uvicorn is a fast ASGI server for Python, often used to run FastAPI and other ASGI-compatible web frameworks. Let’s break it down clearly:

• Uvicorn stands for “Ultra Fast Python Web Server”.
• ASGI server represents to Asynchronous Server Gateway Interface.
• Uvicorn allows us to run running async Python frameworks like FastAPI, Starlette, etc.

Uvicorn is a Python ASGI server, similar to how Apache Tomcat runs Java apps or Node.js serves JavaScript apps — it handles HTTP requests and passes them to your application.

Note: Python is require prior to install Uvicorn as Uvicorn works on top of Python.

# Install Uvicorn
Install Uvicorn: pip install uvicorn

# main → Python file (main.py)
# app → FastAPI instance inside the file
Start Server: uvicorn main:app

# Start Server and automatic reload after change: 
uvicorn main:app - reload

# Specify Hose and Port: 
uvicorn main:app - host 0.0.0.0 - port 8000

# Enable Log: 
uvicorn main:app - reload - log-level info - access-log

# Set Log Level: 
uvicorn main:app - reload - log-level debug

# SSL / HTTPS: 
uvicorn main:app - ssl-keyfile key.pem - ssl-certfile cert.pem - host 0.0.0.0 - port 443

Deployment Commands:

Below are the deployment commands for dev and production:

// Start dev server on port 8000, automatic reload post changes or deployment and set log level as debug.
DEV / QA Command: uvicorn main:app — reload — host 0.0.0.0 — port 8000 — log-level debug

// Start PROD server using Gunicorn with 4 unicorn workers on port 8000, automatic reload post changes or deployment.
PROD Command: gunicorn -k uvicorn.workers.UvicornWorker main:app — workers 4 — bind 0.0.0.0:8000

Note:

• Gunicorn is a WSGI/ASGI server for running Python web apps. Handles process management, load balancing, and concurrent requests.
• -k specifies the worker class as uvicorn.workers.UvicornWorker tells Gunicorn to use Uvicorn as the worker. More workers means better concurrency and CPU utilization.

Worker

Separate process that handles incoming HTTP requests.

• When a request comes in, Gunicorn passes it to one of its workers.
• Each worker can handle one or more requests concurrently (depending on whether it’s sync or async).

Each worker runs its own copy of the FastAPI app.
Number of workers =(2 x CPU cores) + 1
Example: 2 CPU cores → 2*2 + 1 = 5 workers

Note: Adding one worker always is as safey margin.

Work Importance:

• Concurrency → Multiple workers allow handling many requests at the same time.
• Fault tolerance → If one worker crashes, others keep serving requests.

Follow link to read more about Uvicorn.

I hope you found out this article interesting and informative. Please share it with your friends to spread the knowledge.

You can follow me for upcoming blogs follow.
Thank you!