Gunicorn with Uvicorn workers in FastAPI

Introduction

Gunicorn helps run your FastAPI app smoothly by managing multiple workers. Uvicorn workers let Gunicorn handle async FastAPI code efficiently.

You want to run a FastAPI app in production with multiple workers for better performance.

You need to handle many users at the same time without slowing down.

You want to use Gunicorn's process management with FastAPI's async features.

You want to restart workers automatically if they crash.

You want to serve your FastAPI app behind a web server like Nginx.

Syntax

FastAPI

gunicorn -w <number_of_workers> -k uvicorn.workers.UvicornWorker <module_name>:<app_instance>

-w sets how many worker processes Gunicorn runs.

-k uvicorn.workers.UvicornWorker tells Gunicorn to use Uvicorn workers for async support.

Examples

Runs 4 Uvicorn workers from the FastAPI app in main.py named app.

FastAPI

gunicorn -w 4 -k uvicorn.workers.UvicornWorker main:app

Runs 2 workers for the FastAPI app in myapi.py.

FastAPI

gunicorn -w 2 -k uvicorn.workers.UvicornWorker myapi:app

Sample Program

This is a simple FastAPI app that returns a greeting message. You run it with Gunicorn using Uvicorn workers to handle multiple requests efficiently.

FastAPI

from fastapi import FastAPI

app = FastAPI()

@app.get('/')
async def read_root():
    return {"message": "Hello from FastAPI with Gunicorn and Uvicorn workers!"}

# To run this app with Gunicorn and Uvicorn workers, use:
# gunicorn -w 3 -k uvicorn.workers.UvicornWorker main:app

Important Notes

Make sure you install both gunicorn and uvicorn in your environment.

Use an odd number of workers for better load balancing in some cases.

Gunicorn manages worker processes, while Uvicorn handles async requests inside each worker.

Summary

Gunicorn with Uvicorn workers runs FastAPI apps efficiently in production.

Use -w to set worker count and -k uvicorn.workers.UvicornWorker for async support.

This setup improves performance and reliability for real users.

Practice

1. What is the main reason to use Gunicorn with Uvicorn workers when running a FastAPI app?

easy

A. To convert FastAPI apps into desktop applications

B. To run FastAPI apps only in development mode

C. To automatically generate API documentation

D. To efficiently handle multiple requests asynchronously in production

Solution

1. Step 1: Understand Gunicorn and Uvicorn roles

   Gunicorn is a production server that manages multiple worker processes. Uvicorn is an async server that runs FastAPI apps efficiently.

2. Step 2: Combine Gunicorn with Uvicorn workers

   Using Gunicorn with Uvicorn workers allows handling many requests asynchronously and improves performance in production.

3. Final Answer:

   To efficiently handle multiple requests asynchronously in production -> Option D

4. Quick Check:

   Gunicorn + Uvicorn = async production efficiency [OK]

Hint: Gunicorn manages workers; Uvicorn handles async FastAPI [OK]

Common Mistakes:

• Thinking Gunicorn is only for development
• Confusing Uvicorn with API documentation tools
• Believing this setup creates desktop apps

2. Which of the following is the correct command to start a FastAPI app with 4 Uvicorn workers using Gunicorn?

easy

A. gunicorn -k uvicorn.workers.UvicornWorker -w main:app 4

B. uvicorn -w 4 -k gunicorn.workers.GunicornWorker main:app

C. gunicorn -w 4 -k uvicorn.workers.UvicornWorker main:app

D. uvicorn main:app -w 4 -k gunicorn.workers.GunicornWorker

Solution

1. Step 1: Identify Gunicorn command syntax

   Gunicorn uses '-w' to set workers and '-k' to specify worker class.

2. Step 2: Match correct worker class and app reference

   Uvicorn worker class is 'uvicorn.workers.UvicornWorker' and app is referenced as 'main:app'.

3. Final Answer:

   gunicorn -w 4 -k uvicorn.workers.UvicornWorker main:app -> Option C

4. Quick Check:

   Gunicorn -w 4 -k UvicornWorker main:app [OK]

Hint: Gunicorn uses -w for workers and -k for worker class [OK]

Common Mistakes:

• Using uvicorn command instead of gunicorn
• Swapping worker class names incorrectly
• Placing worker count after app name

3. Given this command: gunicorn -w 3 -k uvicorn.workers.UvicornWorker main:app, what will happen when the FastAPI app receives multiple simultaneous requests?

medium

A. Requests will be handled concurrently by 3 async workers improving throughput

B. Only one request will be handled at a time, others will wait

C. The app will crash because Gunicorn cannot run async workers

D. Requests will be queued but processed sequentially by a single worker

Solution

1. Step 1: Understand worker count and async support

   Gunicorn runs 3 workers, each using Uvicorn's async worker class to handle requests asynchronously.

2. Step 2: Effect on request handling

   Multiple requests can be handled concurrently by these 3 async workers, improving throughput and responsiveness.

3. Final Answer:

   Requests will be handled concurrently by 3 async workers improving throughput -> Option A

4. Quick Check:

   3 async workers = concurrent request handling [OK]

Hint: More workers + async = better concurrent handling [OK]

Common Mistakes:

• Assuming only one request at a time
• Thinking Gunicorn can't run async workers
• Believing requests queue without concurrency

4. You run gunicorn -w 2 -k uvicorn.workers.UvicornWorker app:app but get an error: ModuleNotFoundError: No module named 'uvicorn'. What is the likely cause?

medium

A. Uvicorn is not installed in the environment

B. Gunicorn does not support the -k option

C. The app module is missing

D. The worker count -w must be at least 4

Solution

1. Step 1: Analyze the error message

   The error says 'No module named uvicorn', indicating Python cannot find the uvicorn package.

2. Step 2: Check environment setup

   This usually means uvicorn is not installed in the current Python environment where Gunicorn runs.

3. Final Answer:

   Uvicorn is not installed in the environment -> Option A

4. Quick Check:

   Missing uvicorn package causes ModuleNotFoundError [OK]

Hint: Install uvicorn package to fix ModuleNotFoundError [OK]

Common Mistakes:

• Assuming Gunicorn lacks -k support
• Blaming app module instead of missing package
• Thinking worker count causes import errors

5. You want to deploy a FastAPI app with Gunicorn and Uvicorn workers on a server with 8 CPU cores. Which command best balances performance and resource use?

hard

A. gunicorn -w 8 -k uvicorn.workers.UvicornWorker main:app

B. gunicorn -w 4 -k uvicorn.workers.UvicornWorker main:app

C. gunicorn -w 16 -k uvicorn.workers.UvicornWorker main:app

D. gunicorn -w 1 -k uvicorn.workers.UvicornWorker main:app

Solution

1. Step 1: Understand worker count vs CPU cores

   Best practice is to use a moderate number of workers (such as 4) for async apps, but too many workers can cause overhead.

2. Step 2: Choose balanced worker count

   Using 4 workers on 8 cores balances concurrency and resource use without excessive context switching.

3. Final Answer:

   gunicorn -w 4 -k uvicorn.workers.UvicornWorker main:app -> Option B

4. Quick Check:

   4 workers on 8 cores = balanced performance [OK]

Hint: Use fewer workers than CPU cores for async apps [OK]

Common Mistakes:

• Using too many workers causing overhead
• Using only one worker limiting concurrency
• Matching workers exactly to CPU cores without buffer