Processing all queues without having to name them explicitly #5882
-
|
We have a small number of weighted queues defined in our config YAML. Something like this: :queues:
- ['within_30_seconds', 5] # 50% of the time we check this queue
- ['within_5_minutes', 2] # 20% of the time
- ['within_1_hour', 2] # 20% of the time
- ['within_24_hours', 1] # 10% of the timeWe've also defaulted Sidekiq to one of those queues, so we have no In production we have certain machines (Heroku Dynos, but same idea) process certain queues by passing the However, we're now wanting to add some special queues that will only be processes by a certain resource class of machines in production. Basically, a subset of our workload demands A LOT of memory. We still want to maintain our latency-based queues, but have a way to push those memory-hungry workloads to queues which will only be processes by the machines with lot o' memory. We're thinking of going with queue names something like So we'll end up with 8 different queues. 4 of them will run on whatever size machine. And 4 of them ONLY on machines with lots of memory. Using But locally, for developer ergonomics, it'd be really nice to only have to run a single Thank you. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 1 reply
-
|
There’s no way to do this. I’d recommend a Procfile.dev which lists all your queues. |
Beta Was this translation helpful? Give feedback.
There’s no way to do this. I’d recommend a Procfile.dev which lists all your queues.