@noe@sapphic.engineer @julia@eepy.moe also consider to spin up a second redis for cache. that takes quite the load off the queue workers too