The current default RabbitMQCluster configuration does not enable durable queues and auto delete features so nova cannot use them and they are disabled by https://github.com/openstack-k8s-operators/nova-operator/pull/159. Eventually, we want to enable these cluster wide but it needs agreement and adaptation from all the operators using Rabbit.

quorum queues perform significantly better than mirrored queues

https://www.rabbitmq.com/docs/quorum-queues#feature-matrix

when enabling durable queues we should also enable quorum queues

1. rabbit_quorum_queue=true

This was moved from https://github.com/openstack-k8s-operators/nova-operator/issues/160