[Nagiosplug-devel] Suggestion about SNMP errors

[John E. Vincent]

I know this isn't going to thread right because I just signed up for the 
mailing list but:

My opinion on the failure of snmp-agents and honestly all plugin 
failures is that it should return UNKNOWN. This is what the status is 
for. It's up to the individual implementation of Nagios at each site to 
determine if UNKNOWN is an alert condition or not. I know several 
plugins I've seen actually offer an option of how to handle service 
timeouts and such (which would typically be the result from a failed 
agent connection).

Chiel, was there a specific condition or return that you had in mind? 
Maybe an incorrect index on a disk query or such?