Industry standards on RabbitMQ queues and routing keys?

I’m trying to implement a rabbitMQ where I have several different publishers, and several consumers
I am using it as an event bus, so multiple applications are using the same RabbitMQ.

What are industry standards regarding routing keys and queues? I currently have multiple bindings on some queues, which I reuse in my consumer code.
Basically, queue A has 3 bindings 1, 2 and 3. In the consumer code, i listen to the queue, and according to the routing key, I execute one method or another. That sounds weird no?
The problem is that I can’t use the shovel plugin to redrive from the DLQ to the main queue (because I lose the routing key, so I don’t know which method should execute on consumer side).

How are you guys doing? Are you using one routing key per queue? If so, what’s the purpose of routing keys?
Thanks!