From: Jan Schreckenbach <Jan.S<chreckenbach@in...> - 2007-05-28 09:29:27
this weekend my mailbox was flooded by notifications for
services on a Windows host being in status CRITICAL. The
reason was that the Nagios Agent died. I set up a bunch
of service dependencies. Nevertheless some services will
still be reported as CRIT if Nagios agent dies since they
have been checked after the agent dies but before it's
status turned CRIT. This has two disadvantages, first if
I look on the "Service Problems" queue I can not tell were
exactly the problem is if I don't remember the dependencies
(the service may actually be in good state but shown as
CRIT) and secondly if I do some service level reporting a
service should only be reported in critical state if it
really is. The same is true for services being reported
as state OK because they are not checked any longer if
the master service is critical.
I think it might be a good solution to report all dependent
services in state UNKNOWN if the master service is in state
CRTITICAL. Had this ever been discussed and is there a
patch (or any configuration option I'm not aware off)?
Get latest updates about Open Source Projects, Conferences and News.