current topology view
* how many servers
* which agents are connected to which servers
* the server-list that each agent has
group membership event tracking
* when rhq server instance was added to cluster
* when rhq server instance was removed from cluster
balancing heuristics tracking
* repartition tracking
** when and which agents fell over to some backup
** record of actual fallover (from and to server identities)
** why - server membership change, or balancing repartition?
* metric load/throughput on each server in the rhq cluster
control / operations
* remotely shutdown over server instances in the rhq cluster (will only work if not using the embedded agent?)
* put all agents [ for some server instance] into maintenance mode
TEST control / operations
* suspend some server - takes it out of the rhq server cluster, which should initiate an agent repartition
* un-suspend some server - puts some rhq server instance back into the cluster, which should initiate an agent repartition