Broker concurrent limit is invalid

[LL-Ling]

Env

python: 3.12.3
os: window 10
taskiq: 0.11.10

Q

I want to limit the concurrent execution of tasks in the broker, so I set sync_tasks_pool_size=1 and max_async_tasks=1, but it didn't work as expected, and the tasks are still running all at once. I noticed that the Semaphore defined by these two options in the Receiver is not applied in callback and run_task. Is my usage incorrect, or is there a way to limit the number of concurrent tasks? Below is my code. I really like the design of TaskIQ, and I look forward to your response.

Code

import asyncio

from taskiq import InMemoryBroker
from taskiq.kicker import AsyncKicker

from taskiq_example.middleware import TestMiddleware

broker = InMemoryBroker(
sync_tasks_pool_size=1,
max_async_tasks=1
)
broker.add_middlewares(TestMiddleware())

async def best_task_ever(value: int) -> bool:
"""Solve all problems in the world."""
print("execute best_task_ever", value)
await asyncio.sleep(5)
return True

async def main():
await broker.register_task(best_task_ever).kiq(1)
await broker.register_task(best_task_ever).kiq(2)
await broker.register_task(best_task_ever).kiq(3)
await broker.register_task(best_task_ever).kiq(4)

task = broker.register_task(best_task_ever)
await AsyncKicker(
    task.task_name,
    broker,
    task.labels
).with_labels(
    func_value=best_task_ever.__name__,
    func_label="test func"
).kiq(
    value=5
)

await broker.wait_all()

if name == "main":
asyncio.run(main())

Result

It executes 5 tasks concurrently, taking about 5-6 seconds.
I need to control it so that the tasks are executed one by one.