Concurrent task execution in FastAPI - Performance & Optimization

Performance: Concurrent task execution

HIGH IMPACT

Concurrent task execution affects how quickly a FastAPI server can handle multiple requests or background tasks without blocking, improving responsiveness and throughput.

Handling multiple I/O-bound tasks in FastAPI endpoints

FastAPI

from fastapi import FastAPI
import asyncio

app = FastAPI()

@app.get("/async-task")
async def async_task():
    await asyncio.sleep(5)  # non-blocking sleep
    return {"message": "Task done"}

Using async/await allows the event loop to handle other requests while waiting, improving responsiveness.

📈 Performance GainNon-blocking wait enables concurrent handling of multiple requests, reducing input delay.

Handling multiple I/O-bound tasks in FastAPI endpoints

FastAPI

from fastapi import FastAPI
import time

app = FastAPI()

@app.get("/sync-task")
def sync_task():
    time.sleep(5)  # blocking sleep
    return {"message": "Task done"}

The synchronous endpoint runs the blocking sleep in a threadpool, consuming a worker thread and limiting concurrency under high load.

📉 Performance CostConsumes a threadpool thread for 5 seconds, causing queuing and poor scalability.

Performance Comparison

Pattern	DOM Operations	Reflows	Paint Cost	Verdict
Synchronous blocking task	N/A	N/A	N/A	[X] Bad
Asynchronous non-blocking task	N/A	N/A	N/A	[OK] Good

Rendering Pipeline

Concurrent task execution in FastAPI affects the server's event loop and request handling pipeline, enabling multiple tasks to run without blocking each other.

→Request Handling

→Event Loop Scheduling

→Task Execution

⚠️ BottleneckBlocking synchronous calls cause the event loop to stall, delaying all other tasks.

Core Web Vital Affected

INP

Concurrent task execution affects how quickly a FastAPI server can handle multiple requests or background tasks without blocking, improving responsiveness and throughput.

Optimization Tips

1Avoid blocking calls like time.sleep() in async FastAPI endpoints.

2Use async/await for I/O-bound tasks to keep the event loop free.

3Run background tasks asynchronously to improve throughput and responsiveness.

Performance Quiz - 3 Questions

What is the main performance benefit of using async/await in FastAPI endpoints?

A It allows the server to handle other requests while waiting for I/O operations.

B It reduces the size of the server's code bundle.

C It improves the visual layout stability of the page.

D It automatically caches responses to speed up loading.

DevTools: Network and Performance panels

How to check: Use the Network panel to observe request timing and the Performance panel to record event loop activity and task concurrency.

What to look for: Look for long blocking times in the main thread and whether multiple requests are handled concurrently without delay.

Practice

1. What is the main benefit of using asyncio.create_task() in FastAPI?

easy

A. It stops all tasks from running until manually restarted.

B. It makes the code run slower by adding delays.

C. It converts synchronous code into synchronous code.

D. It allows running multiple tasks at the same time without waiting for each to finish.

Solution

1. Step 1: Understand asyncio.create_task() purpose

   This function schedules a coroutine to run concurrently as a separate task without blocking the main program.

2. Step 2: Identify the benefit in FastAPI context

   Using it lets FastAPI handle multiple operations at once, improving responsiveness and efficiency.

3. Final Answer:

   It allows running multiple tasks at the same time without waiting for each to finish. -> Option D

4. Quick Check:

   Concurrent task execution = run tasks simultaneously [OK]

Hint: Remember: create_task starts tasks without waiting [OK]

Common Mistakes:

• Thinking create_task slows code down
• Confusing async with sync code
• Believing tasks stop until restarted

2. Which of the following is the correct way to run two async functions concurrently and wait for both to finish in FastAPI?

easy

A. func1(); func2()

B. await asyncio.create_task(func1()); await asyncio.create_task(func2())

C. await asyncio.gather(func1(), func2())

D. asyncio.run(func1(), func2())

Solution

1. Step 1: Understand how to run and wait for multiple async tasks

   asyncio.gather() runs multiple coroutines concurrently and waits for all to complete.

2. Step 2: Check each option's correctness

   await asyncio.gather(func1(), func2()) correctly uses await asyncio.gather(func1(), func2()). await asyncio.create_task(func1()); await asyncio.create_task(func2()) starts tasks but awaits them separately, not concurrently. func1(); func2() runs functions synchronously. asyncio.run(func1(), func2()) is invalid syntax.

3. Final Answer:

   await asyncio.gather(func1(), func2()) -> Option C

4. Quick Check:

   Use gather to run and wait multiple tasks [OK]

Hint: Use gather to wait all tasks together [OK]

Common Mistakes:

• Awaiting tasks separately losing concurrency
• Calling async functions without await
• Using asyncio.run inside async code

3. Given this FastAPI async endpoint code, what will be the output?

import asyncio
from fastapi import FastAPI
app = FastAPI()

async def task1():
    await asyncio.sleep(1)
    return "Task 1 done"

async def task2():
    await asyncio.sleep(2)
    return "Task 2 done"

@app.get("/run")
async def run_tasks():
    results = await asyncio.gather(task1(), task2())
    return {"results": results}

medium

A. {"results": ["Task 1 done", "Task 2 done"]}

B. {"results": ["Task 2 done", "Task 1 done"]}

C. {"results": ["Task 1 done"]}

D. TimeoutError

Solution

1. Step 1: Analyze the async tasks and their sleep durations

   task1 sleeps 1 second, task2 sleeps 2 seconds. Both run concurrently with gather.

2. Step 2: Understand gather result order

   asyncio.gather returns results in the order of the input tasks, regardless of completion time.

3. Final Answer:

   {"results": ["Task 1 done", "Task 2 done"]} -> Option A

4. Quick Check:

   Gather preserves input order in results [OK]

Hint: Gather returns results in input order, not finish order [OK]

Common Mistakes:

• Assuming results order matches finish order
• Expecting partial results if one task is slower
• Confusing gather with wait

4. What is wrong with this FastAPI code snippet for running tasks concurrently?

import asyncio
from fastapi import FastAPI
app = FastAPI()

async def task():
    await asyncio.sleep(1)
    return "Done"

@app.get("/test")
async def test():
    task1 = asyncio.create_task(task())
    task2 = asyncio.create_task(task())
    return {"result": task1.result() + task2.result()}

medium

A. Calling task1.result() before task completion causes an error.

B. Using create_task is not allowed in FastAPI endpoints.

C. The tasks should not be async functions.

D. The endpoint should not be async.

Solution

1. Step 1: Identify how task results are accessed

   Calling task1.result() immediately after creating the task tries to get the result before the task finishes, causing an error.

2. Step 2: Understand correct way to wait for task completion

   You must await the task or use await asyncio.gather(task1, task2) before accessing results.

3. Final Answer:

   Calling task1.result() before task completion causes an error. -> Option A

4. Quick Check:

   Access task result only after await [OK]

Hint: Always await tasks before accessing results [OK]

Common Mistakes:

• Accessing .result() too early
• Thinking create_task is forbidden
• Making endpoint synchronous

5. You want to run three independent async tasks in a FastAPI endpoint and return their results as soon as all finish. Which approach correctly handles errors in any task without stopping others and returns all results or error messages?

hard

A. Use asyncio.create_task() for each task but do not await them before returning.

B. Use asyncio.gather(task1(), task2(), task3(), return_exceptions=True) and process results for exceptions.

C. Use await task1(); await task2(); await task3() sequentially and catch exceptions globally.

D. Run tasks synchronously in a loop and catch exceptions inside each.

Solution

1. Step 1: Understand error handling in concurrent tasks

   Using asyncio.gather() with return_exceptions=True lets all tasks run concurrently and collects exceptions as results without stopping others.

2. Step 2: Compare options for correct concurrency and error handling

   Use asyncio.gather(task1(), task2(), task3(), return_exceptions=True) and process results for exceptions. correctly runs all tasks concurrently and handles errors individually. Use await task1(); await task2(); await task3() sequentially and catch exceptions globally. runs tasks sequentially, losing concurrency. Use asyncio.create_task() for each task but do not await them before returning. does not await tasks, so results are unavailable. Run tasks synchronously in a loop and catch exceptions inside each. runs synchronously, blocking concurrency.

3. Final Answer:

   Use asyncio.gather(task1(), task2(), task3(), return_exceptions=True) and process results for exceptions. -> Option B

4. Quick Check:

   Gather with return_exceptions handles errors concurrently [OK]

Hint: Use gather with return_exceptions=True for safe concurrency [OK]

Common Mistakes:

• Awaiting tasks sequentially losing concurrency
• Not awaiting tasks before returning
• Ignoring exceptions stopping all tasks