| survivor: Obscure Tasks Explained | |||||||||||||
Some useful tasks are perhaps not as obvious as they could be. This document explains some examples.
In order to test configuration files before notifying the scheduler, use the command line interface. % sc checkcf Configuration file parse successful
To monitor workstations that might turn off, try one of these approaches:
Making a test dependent on two checks is simple: define the test as a composite check.
Response times (execution duration) are automatically recorded for each executed check. To examine these times, use the responsetime report module. To automatically monitor response times via a check, use the report check module. A configuration like this one might be useful:
check responsetime {
module report {
module responsetime
instance myinstance
service http
source check
recordspan tnt[3600] # only look at the last hour
warn gt[1000] # warn if response time is > 1 second
problem gt[2000] # problem if response time is > 2 seconds
}
}
Flapping can be detected with the flap report module. To automatically monitor response flapping via a check, use the report check module. A configuration like this one might be useful:
check responsetime {
module report {
module flap
instance myinstance
service ping
source check
recordspan tnt[86400] # only look at the last day
warn gt[4] # warn if > 4 flaps (eg: ok->bad->ok->bad->ok)
problem gt[8] # problem if > 8 flaps
}
}
$Date: 2007/03/29 12:26:16 $ $Revision: 1.1 $ |
keywords |
||||||||||||