Is this behaviour constant ?
Or does it just happen from time to time ?
If it just happens sometimes ... well, I can't help you much: I have the same problem with a few of my nagios servers: sometimes (one or twice per month), the just stop doing their checks. Everything looks normal (processes present, web interface up and running) except that nagios is doing nothing
=> the "solution": we had to implement a "Heartbeat" test which is sent (by SNMP) to Netcool, which raises an alert if 2 heartbeats are absent....
Sorry that my answer doesn't help much