A basic goal of performing fault management on any system is the ability to find out easily the collective alarm status of the system. In simple terms, you can go to an "agent" and ask for a list of the active system alarms. This answers the question "what is wrong with my system now".

Currently in jboss, there is no clear distinction what constitutes an alarm, how you can produce or map alarms, and where you can get them. However, there is a lot of management functionality and transient events exposed through jmx operations/attributes/notifications. So we can try to introduce a couple of components that will gather this information into a single location.

We'll introduce those component as subtasks of this case.