I have a celery
app consuming from large amount of queues with gevent
concurrency.
For each queue I need to have some guaranteed number of workers, but the number is different for each queue. For example:
- qA must have 10
- qB must have 25
- qN must have M etc.
I'd like to avoid to start dedicated celeryd
for each queue due to the large total amount of queues.
Workers do a lot of IO and I use -P gevent
so don't care how many subworkers celeryd
starts.
Is there a way to configure a single celeryd
instance to assign some specific number of workers to different queues?