You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A feature that would be useful to us is to put a limit on the combined size of all the pools in a group. Our use case is kubernetes pods - we only want to run up to N of them overall, but we have a pool of pods for each image that we run. It would be nice if each pool gets its init_size guaranteed, but can only grow if the total number of workers across all pools in the group has not exceeded N. That way a large workload that only uses one image could run something close to N workers, and when that workload is finished there would be more slots avialable for pods that use other images.
Just sketching the idea here as I've yet to try out Pooler - we're currently using Poolex which lacks the grace period to retire extra workers - it's that feature that has me looking into Pooler.
The text was updated successfully, but these errors were encountered:
A feature that would be useful to us is to put a limit on the combined size of all the pools in a group. Our use case is kubernetes pods - we only want to run up to N of them overall, but we have a pool of pods for each image that we run. It would be nice if each pool gets its
init_size
guaranteed, but can only grow if the total number of workers across all pools in the group has not exceeded N. That way a large workload that only uses one image could run something close to N workers, and when that workload is finished there would be more slots avialable for pods that use other images.Just sketching the idea here as I've yet to try out Pooler - we're currently using Poolex which lacks the grace period to retire extra workers - it's that feature that has me looking into Pooler.
The text was updated successfully, but these errors were encountered: