Custom middleware creation in FastAPI - Performance & Optimization

Performance: Custom middleware creation

MEDIUM IMPACT

This affects the request-response cycle speed and server throughput by adding processing steps before and after handling requests.

Adding custom logic to process requests and responses

FastAPI

from fastapi import FastAPI
from starlette.middleware.base import BaseHTTPMiddleware
import asyncio

app = FastAPI()

class EfficientMiddleware(BaseHTTPMiddleware):
    async def dispatch(self, request, call_next):
        # Non-blocking async processing
        await asyncio.sleep(0.1)  # does not block event loop
        response = await call_next(request)
        return response

app.add_middleware(EfficientMiddleware)

Uses async sleep to avoid blocking event loop, allowing other requests to be processed concurrently.

📈 Performance GainNon-blocking delay improves throughput and reduces request latency under load.

Adding custom logic to process requests and responses

FastAPI

from fastapi import FastAPI
from starlette.middleware.base import BaseHTTPMiddleware

app = FastAPI()

class SlowMiddleware(BaseHTTPMiddleware):
    async def dispatch(self, request, call_next):
        # Heavy synchronous processing
        import time
        time.sleep(0.1)  # blocks event loop
        response = await call_next(request)
        return response

app.add_middleware(SlowMiddleware)

Blocking the event loop with synchronous sleep delays all requests, reducing throughput and increasing latency.

📉 Performance CostBlocks event loop for 100ms per request, increasing response time and reducing concurrency.

Performance Comparison

Pattern	DOM Operations	Reflows	Paint Cost	Verdict
Blocking synchronous middleware	N/A	N/A	N/A	[X] Bad
Async non-blocking middleware	N/A	N/A	N/A	[OK] Good
CPU-heavy synchronous middleware	N/A	N/A	N/A	[X] Bad
CPU work offloaded to thread pool	N/A	N/A	N/A	[OK] Good

Rendering Pipeline

Middleware runs during the server's request-response cycle before and after the main handler. It adds processing steps that can delay response generation.

→Request Handling

→Response Generation

⚠️ BottleneckSynchronous or CPU-heavy middleware blocks the event loop, delaying all requests.

Optimization Tips

1Avoid synchronous blocking code in middleware to keep the event loop free.

2Use async/await and non-blocking operations inside middleware.

3Offload CPU-intensive tasks to threads or background workers.

Performance Quiz - 3 Questions

What is the main performance risk of using synchronous blocking code in FastAPI middleware?

A It blocks the event loop, increasing request latency and reducing throughput.

B It increases the size of the response payload.

C It causes layout shifts in the browser.

D It improves CPU utilization.

DevTools: Performance

How to check: Record a server request in DevTools Performance panel, look for long tasks blocking the main thread or event loop.

What to look for: Long blocking tasks or delays before response indicate middleware blocking; short async tasks indicate good performance.