python - Can Celery deliver task results in under 10ms? -

python - Can Celery deliver task results in under 10ms? -

i trying execute task on celery , block on it's result less 10ms overhead.

a self contained illustration (tasks.py):

from celery import celery app = celery('tasks', broker='librabbitmq://guest@localhost//', backend='amqp') @app.task def add(x, y): homecoming x + y

test script:

from timeit import timeit tasks import add together def f(): add.delay(2, 2).get() timeit(f, number=100)

this runs 40ms per request on i5 8gb ram debian linux machine.

imho redis result backend in celery relies on polling, isn't ideal low latency requirements?

the solution need deliver under 10ms rtt overhead 4 machines adding tasks every second, , 2 task workers executing tasks.

python performance rabbitmq celery low-latency