Python FastAPI: Slow Response Time with Large JSON Payloads in Asynchronous Endpoints
I've been researching this but I'm working on a personal project and I'm updating my dependencies and I've been banging my head against this for hours. I'm experiencing significant performance issues with my FastAPI application when handling large JSON payloads in asynchronous endpoints. The application is designed to process requests that include large lists of nested dictionaries. For instance, when I send a payload of approximately 10,000 items, the response time increases dramatically. I've noticed that the endpoint that processes this data can take up to 10 seconds to respond, which is unacceptable for a real-time application. Here’s a simplified version of my endpoint: ```python from fastapi import FastAPI from pydantic import BaseModel from typing import List app = FastAPI() class Item(BaseModel): id: int name: str details: str class RequestPayload(BaseModel): items: List[Item] @app.post('/process-items/') async def process_items(payload: RequestPayload): # Simulate processing time processed_items = [f"{item.name} processed" for item in payload.items] return {"status": "success", "processed_items": processed_items} ``` I have tried several optimizations, such as using `async` for the function and limiting the size of requests, but the performance issue persists. My FastAPI version is 0.68.0, and I'm running this on an ASGI server with `uvicorn`. Additionally, I have also tested using `httpx` for making requests and verified that the payload size is consistent, but the processing time keeps scaling poorly with the number of items. I would appreciate any advice on improving the performance of this endpoint or techniques to handle large payloads more efficiently. I'm working on a CLI tool that needs to handle this. For context: I'm using Python on Linux. This issue appeared after updating to Python latest. What would be the recommended way to handle this? Hoping someone can shed some light on this. This issue appeared after updating to Python 3.10. Any suggestions would be helpful.