Proper alert escalations question

[Charles Jones]

The goal here is for the secondary on-call person to have a restful 
nights sleep, assuming that the primary oncall person fixes the problem 
that caused the alert.

* The primary should be alerted immediately upon a problem
* The primary should get a RECOVERY page when the alert recovers
* The secondary should ONLY get an alert if something is red for 30+ minutes
* The secondary should get a RECOVERY page, ONLY if they were alerted in 
the first place (for a +30min event).
* If the primary ACKs an alert, the secondary will not be emailed unless 
the ACK expires and the service is still red.

If I use a ruleset of:

 HOST=www.foo.com SERVICE=http
      MAIL primary_oncall at foo.com  FORMAT=sms COLOR=red RECOVERED
      MAIL secondary_oncall at foo.com FORMAT=sms COLOR=red DURATION>30 
RECOVERED

I do not believe this will work? Because the secondary will still get 
woke up by the RECOVERED message? Or is Hobbit smart enough that it only 
sends RECOVERED messages to only MAIL recipients that have previously 
received an alert?