Steve,
I implemented this almost generically - except for the following changes:
1) I had an e-mail sent to me when it occurred. This will help determine if something is going on and not being properly monitored or managed.
2) I also have a 'console setup' rexx which is called when this occurs. For some specific consoles in operations, this rexx changes the RTME, RNUM, ROUT and MFORM to our standards. I've seen cases where MFORM is changed (dropping the time from the message - which affects some other (Non-CA) software) or roll time is increased to multiple seconds. So, I have automation check/fix these for 2 specific consoles on each system.
3) I changed the numbuf minimum from 0 to 25. I found that numbuf is often a small number on our systems even though things are working ok. A different console would be the one with a backlog of messages. This isn't too big a problem since it can be called again in 10 seconds.
The newest (12.1) version appears different from the 11.9 version. It will probably be just fine as well...maybe I should consider upgrading.