Question

An AI wrote this FastAPI endpoint that calls a payments provider. Latency looked fine in a single-user test, but under modest concurrency the whole service's p99 spikes and unrelated endpoints stall. Find the bug and explain why it only shows up under load.

1import requests2 3@app.get("/quote/{symbol}")4async def get_quote(symbol: str):5    resp = requests.get(f"https://prices.internal/v1/{symbol}", timeout=5)6    price = resp.json()["price"]7    return {"symbol": symbol, "price": price}