D.1. ActiveMQ endpoint discovery
If you were to deploy ActiveMQ brokers into a collection of standalone containers, you would have to provide an explicit list of broker locations to any messaging clients. In the case of a cloud deployment, where servers are frequently shut down and restarted in different locations, this type of configuration is very difficult to maintain.
Deploying the ActiveMQ brokers on a fabric, however, is much easier to manage, because you can exploit the Fabric discovery mechanism, which enables location transparency. When they start up, the brokers register themselves with the Fabric registry. Clients can then consult the Fabric registry to discover brokers. Moreover, because the discovery mechanism automatically detects when brokers fail and when new brokers start, there is also support for real-time discovery and load balancing.
For example, in the context of a Fabric deployment, it is typically sufficient to configure clients with a discovery URL like the following: