Feature Request: Support for error queues for rabbitmq

[jonsalvas]

Now the reactive-messaging-rabbitmq plugin allows me configure queues and bind a channel to it in order to consume the messages from it. It is required to have an @incoming channel and consume the messages, otherwise the configuration will ignore that you add an incoming channel with a queue in the application.yaml configuration.

However there are cases where it makes sense that the application creates queues that are not consumed. The best use-case IMO is the error-queue which is the last resort if after x retries from the the dead-letter-queue (with timeout etc. ) the message could not be processed for whatever reason. In this case it will simply be put to the error queue through an error exchange.

Therefore it would be nice to have the option to define custom queues as otherwise the queue has to be declared manually (or by infrastructure scripts) before the app is deployed which is IMO against the cloud idea and the idea of exchange consumers in rabbitmq.

What do you think?

Implementation ideas:

add a config parameter in mp.messaging.incoming.*.createOnStart which allows defining a queue in the application configuration without the need of consuming it.

OR

add configuration outgoing.*.queue.declare in order to declare outgoing queues without consuming them