Nagios Support Forum

We are implementing a distributed monitoring environment with approx. 40 customer environments using over 30 XI servers and NagiosFusion.
In several of these environments we have "special needs" hosts.
There are a variety of scripts running in these "special needs" environments.
They follow this structure:

Scripts are executed locally on "special" hosts (for instance VMS and DEC)
The results are passed to a Kiwi Syslog server in that environment.
The Kiwi server interprets these results and generates a windows event in Application Log with event-id:
777 for Informational
888 for warning
999 for critical.

We started out monitoring the eventlogs, but that caused problems.
(for instance check-interval 10 minutes. If no new events occur within the last 10 minutes status is green (OK) regardless if previous errors are fixed ornot)
(for instance check-interval 24 hours. If errors are fixed the status will remain critical because it still is an occurence within the last 24 hours.)

Because of this we decided to have Kiwi send out SNMP traps.(See attachment) Because every SNMP trap is an event and can be handled individually.
(In the future we would have the hosts trap directly to Nagios, but that is a lot of scripting changes and therefore outside of the scope of the implementation project)

On thing with traps and snmptt specifically, the oids must match exactly. So your snmptt.conf is looking for .1.3.6.1.4.1.20580.69.0.1 however your kiwi server is sending as .1.3.6.1.4.1.20580.69. This can be easily resolved on either end, adding the .0.1 to the kiwi server enterprise oid line, or removing those from the snmptt.conf file and doing a service restart on snmptt should get those coming in properly. Give either suggestion a try, but not both, and let us know how those changes work for you.

Thank you!!

It works. The traps are now linked to the host.
I guess i should have seen that error myself.

This helped build my understanding of snmp trapping, but unfortunately does not solve my original problem.
I guess i will have to look at an alternative solution.

Thanks for all the help.

Regards,

Rik Lijkendijk