WSGI servers (Gunicorn, uWSGI) in Flask - Performance & Optimization

Performance: WSGI servers (Gunicorn, uWSGI)

HIGH IMPACT

This affects how fast your Flask app starts serving requests and handles concurrent users, impacting server response time and throughput.

Serving a Flask app to multiple users efficiently

Flask

gunicorn -w 4 -b 0.0.0.0:8000 myapp:app  # Using Gunicorn with multiple workers

Gunicorn runs multiple worker processes to handle many requests concurrently, improving throughput and responsiveness.

📈 Performance Gainhandles multiple requests in parallel, reducing request wait time and improving INP

Serving a Flask app to multiple users efficiently

Flask

app.run()  # Using Flask's built-in server in production

Flask's built-in server is single-threaded and not designed for production, causing slow response and blocking under load.

📉 Performance Costblocks requests under concurrent load, leading to high latency and poor INP

Performance Comparison

Pattern	Concurrency	Request Latency	Resource Usage	Verdict
Flask built-in server	Single-threaded	High under load	Low but inefficient	[X] Bad
Gunicorn with multiple workers	Multi-process	Low	Moderate CPU/memory	[OK] Good
uWSGI single process/thread	Single-threaded	High latency	Low but blocking	[X] Bad
uWSGI multiple processes/threads	Multi-process/thread	Low latency	Higher resource use but efficient	[OK] Good

Rendering Pipeline

WSGI servers receive HTTP requests, pass them to the Flask app, then send back responses. Efficient concurrency reduces waiting time before response generation.

→Request Handling

→Application Processing

→Response Sending

⚠️ BottleneckRequest Handling when concurrency is low or server is single-threaded

Core Web Vital Affected

INP

This affects how fast your Flask app starts serving requests and handles concurrent users, impacting server response time and throughput.

Optimization Tips

1Never use Flask's built-in server for production; use a WSGI server like Gunicorn or uWSGI.

2Configure multiple worker processes and threads to handle concurrent requests efficiently.

3Monitor server response times under load to adjust concurrency settings for best performance.

Performance Quiz - 3 Questions

Why is using Flask's built-in server in production a bad idea for performance?

A It is single-threaded and blocks requests under load

B It uses too much memory

C It does not support HTTP

D It automatically caches responses

DevTools: Network and Performance panels

How to check: Use Network panel to measure response times under load; use Performance panel to record request handling delays and CPU usage.

What to look for: Look for consistent low response times and no long blocking periods indicating good concurrency and server responsiveness.

Practice

1. What is the main role of a WSGI server like Gunicorn or uWSGI when running a Flask application?

easy

A. To connect the Flask app to the web and handle multiple user requests efficiently

B. To write the Flask application code

C. To create the HTML templates for the Flask app

D. To manage the database connections directly

Solution

1. Step 1: Understand WSGI server purpose

   WSGI servers act as a bridge between the web and Flask app, managing requests.

2. Step 2: Identify what Gunicorn and uWSGI do

   They handle multiple users and serve the Flask app efficiently in production.

3. Final Answer:

   To connect the Flask app to the web and handle multiple user requests efficiently -> Option A

4. Quick Check:

   WSGI server role = Connect app and handle requests [OK]

Hint: WSGI servers link web and app, not write code [OK]

Common Mistakes:

• Thinking WSGI servers write app code
• Confusing WSGI with template engines
• Assuming WSGI manages databases

2. Which of the following is the correct command to start a Flask app using Gunicorn on port 8000?

easy

A. gunicorn app.py --port 8000

B. gunicorn run app --port=8000

C. gunicorn -w 4 -b 0.0.0.0:8000 app:app

D. gunicorn app:app --workers=2 --port=8000

Solution

1. Step 1: Recall Gunicorn syntax

   Gunicorn uses '-w' for workers and '-b' for binding address and port.

2. Step 2: Check correct app reference and port

   Correct format is 'gunicorn -w 4 -b 0.0.0.0:8000 app:app' to run app with 4 workers on port 8000.

3. Final Answer:

   gunicorn -w 4 -b 0.0.0.0:8000 app:app -> Option C

4. Quick Check:

   Gunicorn uses -w and -b flags [OK]

Hint: Gunicorn uses -w for workers and -b for bind address [OK]

Common Mistakes:

• Using --port instead of -b for binding
• Incorrect app reference like app.py instead of app:app
• Wrong flag order or missing workers

3. Given the command gunicorn -w 3 -b 127.0.0.1:5000 app:app, what will happen when you try to access the Flask app from another computer on the network?

medium

A. The app will be accessible from any computer on the network

B. The app will automatically bind to all interfaces

C. The app will refuse all connections and crash

D. The app will only be accessible from the local machine (127.0.0.1)

Solution

1. Step 1: Understand binding address 127.0.0.1

   Binding to 127.0.0.1 means the server listens only on localhost, not external network.

2. Step 2: Check network accessibility

   Other computers cannot access the app because it is not bound to the external IP or 0.0.0.0.

3. Final Answer:

   The app will only be accessible from the local machine (127.0.0.1) -> Option D

4. Quick Check:

   127.0.0.1 = localhost only [OK]

Hint: 127.0.0.1 binds only local access, not network [OK]

Common Mistakes:

• Assuming 127.0.0.1 allows network access
• Thinking Gunicorn auto-binds to all interfaces
• Confusing localhost with external IP

4. You run uwsgi --http :5000 --module app:app but get an error saying the module 'app' cannot be found. What is the most likely cause?

medium

A. The app.py file is missing or not in the current directory

B. The uWSGI command requires the --wsgi-file option instead of --module

C. The port 5000 is already in use by another process

D. The Flask app is not installed properly

Solution

1. Step 1: Understand module loading in uWSGI

   Using --module expects a Python module named 'app' with an 'app' variable inside.

2. Step 2: Check file presence and location

   If 'app.py' is missing or not in the current directory, uWSGI cannot find the module.

3. Final Answer:

   The app.py file is missing or not in the current directory -> Option A

4. Quick Check:

   Missing app.py causes module not found error [OK]

Hint: Check app.py file exists in current folder [OK]

Common Mistakes:

• Confusing --module with --wsgi-file
• Assuming port conflicts cause module errors
• Ignoring file location issues

5. You want to deploy a Flask app with uWSGI to handle 8 worker processes and bind it to all network interfaces on port 8080. Which command correctly achieves this?

hard

A. uwsgi --http 0.0.0.0:8080 --workers 8 --module app:app

B. uwsgi --http :8080 --processes 8 --module app:app --master

C. uwsgi --socket 0.0.0.0:8080 --processes 8 --wsgi-file app.py

D. uwsgi --http 127.0.0.1:8080 --workers 8 --module app:app

Solution

1. Step 1: Identify correct uWSGI flags for HTTP and workers

   --http :8080 binds to all interfaces on port 8080; --processes 8 sets 8 worker processes.

2. Step 2: Confirm module and master process usage

   --module app:app correctly points to the Flask app; --master enables master process for better management.

3. Step 3: Eliminate incorrect options

   A: uwsgi --http 0.0.0.0:8080 --workers 8 --module app:app uses invalid --workers flag (use --processes); B: uwsgi --http 127.0.0.1:8080 --workers 8 --module app:app binds only to localhost; C: uwsgi --socket 0.0.0.0:8080 --processes 8 --wsgi-file app.py uses --socket which is for socket mode, not HTTP.

4. Final Answer:

   uwsgi --http :8080 --processes 8 --module app:app --master -> Option B

5. Quick Check:

   Use --http :port and --processes with --master for production [OK]

Hint: Use --http :port and --processes with --master for uWSGI [OK]

Common Mistakes:

• Using --workers instead of --processes
• Confusing --socket with --http
• Binding only to 127.0.0.1 for network apps