The best way (when possible) is to monitor the servers the same way clients query the HA cluster. Consider you have ha-cluster with 2 nodes (nodeA and nodeB), you'll setup the monitoring to query ha-cluster, rather than nodeA or nodeB. The clustering mechanism will direct the monitoring queries to either nodeA or nodeB, depending on which one is active.
Hope this helps!
Thanks a lot for the reply. I guess I need a bit more help on how to go about configuring a Patrol Agent to monitor as a cluster. The servers are not in a dedicated cluster but have a rudimentary configuration that starts up the processes on "NodeB" should they stop on "NodeA". How do I get the Patrol Agent to recognise that behaviour?