I'm running a prisma cluster (1.33.0) on k8s. Everything works per docs but I noticed that prisma tacks on the pod name as part of the rabbitmq queue name, and as the prisma pods get restarted in the k8s cluster the rabbitmq prisma queues (project-schema-invalidation-exchange and sss-events-exchange) that backed the deleted pods remain and are actually still listed as having consumers so the queues never auto-delete. Anyone running with this type of setup that can verify if this is indeed a bug?