Add container/host identifier for messages sent via queue #413

There should be a way to identify which container/host instance of the API is sending messages that fail to be processed by the federation consumer so that if one specific instance seems to be having issues it can be restarted/killed/whatever to try to resolve problems (or pulled out of rotation and investigated)