#celery

ah asksol, hi

thank you

let me catch up

asksol: ideally i'd like to say "every worker is running and we verified roundtrip success with rabbit"

asksol: what aspect of my description makes it seem like i have two workers with the same name?

reading the configs now, none of these workers actually specify a name, only the -Q they are consuming from

ping() is returning correctly btw. i was just hoping to have a greater assurance that everything is going smoothly, since one of my workers could have failed to start but i think ping() would still respond happily

and congrats on kombu 4.0.0!

the worker could definitely respond to ping and not be functional, you could send it an actual task, but then if the queues are congested...

heh that's my concern precisely

a lot of these queues are pretty hairy-- lots of long running tasks

asksol: ping() can't succeed with no broker tho right?

i might have to settle with that for now

is there an idiom to prevent running tasks from queue X while the size of queue Y is >= N

meltingwax: what do you expect to happen with those messages?

meltingwax: stay in queue X until queue Y is <N

meltingwax: rabbitmq has a max-length, but I'm not aware of a min-length

if max length is exceeded, the messages are dropped afaik

the *new* messages