RabbitMq – Should I dynamically declare queues and make one consumer listen to all queues?

I have application X that can create “topics” where users can do operations, I want each “topic” to be a new “queue”, so every “topic” operations are independent from another. I can create new queue on the moment of “topic” creation from this application.

What is the optimal solution for the receiving application Y – if one consumer should listen to all queues how do I create new queue dynamically while the consumer is already listening to the previous queues? Or should every queue gets its own consumer (maybe via API call to application Y to start consuming the new queue)?