[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: [hobbit] DURATION explanation in hobbit-alerts.cfg



Actually, I think the alert model is based on the "state" that a
"service" is in.  So what you might alert on is that the "conn" service
has been "red" for at least one minute.  This may or may not be very
effective, given the built in functions in the Hobbit ping testing.  If
you want to know that something has failed two successive poll cycles,
and the poll cycle is 5 minutes, you could set DURATION to > 6.
 
GLH


________________________________

	From: Taylor Lewick [mailto:tlewick (at) tradebotsystems.com] 
	Sent: Friday, April 11, 2008 1:17 PM
	To: hobbit (at) hswn.dk
	Subject: [hobbit] DURATION explanation in hobbit-alerts.cfg
	
	

	Okay, I've read the hobbit-alerts man page and docs.

	 

	Under Duration the man page says: Rule matching an alert if the
event has lasted longer/shorter than the given duration.  And then it
goes on to give examples and those are in minutes.

	 

	But for the ping test for a host being down, I only want to send
the alert if a host is unresponsive for at least 1 minute (trying to
eliminate a bunch of false positives)  So I set it up for a single host
with the following rule...

	 

	HOST=test103

	MAIL xyz DURATION>1 REPEAT=10 COLOR=red RECOVERED

	 

	So based on the alerts I've received (see below), Duration isn't
per minute.  So is there a way to set the alerts so I can specify don't
send an alert unless its 1 minute.

	 

	Service conn on test103 is not OK : Host does not respond to
ping

	System unreachable for 2 poll periods (19 seconds)

	&red test103 is unreachable