Hi Wayne,
In our env, we have monitors setup for each component - masters, CMs and the Fault Monitor.
Our oncall is notified if any component goes down.
Knowing that if the FM goes down the master will still process jobs makes this a sufficient solution for our needs.
If the primary master and FM went down simultaneosly then it's most likely a much larger issue which failing over wouldn't fix anyway.
Another option, if you have a high availablility cluster that can be setup transparently to the FM you could test it there and see if it works well for you. I'd open a case to verify if it's supported it or not by the BU.
-Prakash
