f96fc6d1-3b42-4079-bec9-6c8bbc962296

what specifically are you confused on that guide, id be happy to explain something you are stuck on in more depth

I've set up the Redis instance, and then made a celery_app.py

from celery import Celery
from dotenv import load_dotenv
import os

load_dotenv()

def make_celery(app_name=__name__):
    return Celery(
        app_name,
        broker=os.environ['REDIS_URL'],
        backend=os.environ['REDIS_URL']
    )

celery = make_celery()

I kinda just got this code online, haven't used celery before

what command do you run locally to start celery?

I was just running using python, that's how I run my flask dev server

would I be correct in assuming you have celery running locally just fine?

well this code all it does is just run and then stops, so no errors but not like it stays open

So that's kind of what I'm unsure of

would I also need to create another flask server to incorporate the celery service? but I guess that doesn't make sense

nope, you'd have two services, one that runs only flask, and one that runs only celery

but please get this working locally first, then once you got it working locally, I'll be here to help you run it on railway!

cool

Ok got it working on local

Now trying to put it onto railway, and im getting this:

invalid type: string "celery -A tasks worker --loglevel=info", expected a map

that's the start command I'm using, or should I use this one:

celery -A liftoff worker -l info --concurrency=3

But then what would liftoff be?

What command do you use locally to run just celery

celery -A tasks worker --loglevel=info

This one seems to work, so using it

instead of a procfile, set the command in the service settings

Just did, should I delete the procfile?

yep

you also need to attach a branch in the service settings

Ok so done that, delete procfile and then added the command

Also I noticed, when I push to github it's not pushing a new version onto railway with the /worker folder

But the /app folder is pushign just fine when there is a change

thats because you didnt have a branch connected up until a few minutes ago, that issue should be sorted out now

when you run the celery command locally, do you do that from within the worker folder?

yeah I do

can you delete all the __pycache__ folders from your repo and then make sure they are being gitignored

Deleted it and added to git ignore

so now the issue is that isnt a typical python project layout, despite there being python files, nixpacks doesnt know that you want it to run python

post the folder structure of the app, and the contents of the nixpacks?

so add a requirements.txt file with all the dependencies you need, and we will see if that helps

folder structure is as printed in that screenshot, and there is no nixpacks.toml file

ok i believe this has worked

sweet

awesome ty, now just need to use it from my flask service

now, you will absolutely need to set the concurrency, the default values will use a lot of memory

ok let me do that now

Any recommendations? or how should I know what to set it to

set it to the numder of jobs you could have running at the same time, if you will only ever run one job at any one time, set it to 1

would it be something like

celery -A app.celery_app worker --concurrency=1 --pool=prefork --loglevel=info --events

what would happen if i try to run a job back to back? does that count as just 1?

ty

as long as they run back to back you're fine, since thats still 1 job at any one point in time

ok cool

I think it will do it one at a time, if you set the concurrency to one.

And it will do it depending on whats in the redis

see the difference?

oh yeah a lot nicer

awsome, any other questions?

that should be it for now!

thank you

no problem, ill mark this as solved, but feel free to open another thread if you need further help!

!s