So, it's increasing in few hours or days - even when the worker is not processing any jobs.
We only run 6 workers.
Those workers always killed by OS automatically when its memory usage keeps increasing.
Is there anyone here who's already fix it?
Because I also see many people reports this issue but it looks like there's no final conclusion yet.
=> the guy moved to sidekiq. We planned this too but our codes already coupled with delayed_job heavily. Moving to sidekiq will be really a huge job for us.
=> tried some solutions there but doesn't work for our case.