Sep 30, 2009 3:27 PM
Dealing With Exit Codes
-
Like (0)
Greetings,
Recently I've been working on a project that involves migrating off of our current monitoring system (Nagios) to zenoss. We currently have numerous checks that have been coded in house, and return the following values via SNMP:
I'm designing a template that will create events when one of these conditions exist and based on those events an alert will sound if the exit code does not equal 0. Currently I have a data source setup to pole certain SNMP OIDs, but I run into an issue when creating the thresholds.
The only feasible way I'm able to create critical/warning/unknown events I see are to create thresholds based setting the min/max values on the following:
Is there anyway to have a threshold be reached if the datasource is equal to a value? The problem with this setup is that it invalidates the credibility of having different event severity. For example, if a datasource of a SNMP check reaches a warning state (the value of the datasource would equal 1), then a warning event will be created, which is ok. On the other hand if a datasource is in an unknown state (the value of the datasources would equal 3) then three events get created (one for each threshold that has been exceeded).
Has anyone has any issues with this?
You should consider the point threshold Zenpack I think.
--
James Pulver
Information Technology Area Supervisor
LEPP Computer Group
Cornell University
bryan.yurecka wrote, On 9/30/2009 3:28 PM:
Greetings,
Recently I've been working on a project that involves migrating off of our current monitoring system (Nagios) to zenoss. We currently have numerous checks that have been coded in house, and return the following values via SNMP:
0 - meaning everything is OK
1 - meaning that a condition that may be of some warning exists
2 - meaning a critical condition exists
3 - meaning the status is unknown and something needs to be reconfigured inorder for the check to function properly
I'm designing a template that will create events when one of these conditions exist and based on those events an alert will sound if the exit code does not equal 0. Currently I have a data source setup to pole certain SNMP OIDs, but I run into an issue when creating the thresholds.
The only feasible way I'm able to create critical/warning/unknown events I see are to create thresholds based setting the min/max values on the following:
by setting the Min value 0, and Max to 0 for warning
by setting the Min value 0, and Max to 1 for critical
by setting the Min value 0, and Max to 2 for unknown
Is there anyway to have a threshold be reached if the datasource is equal to a value? The problem with this setup is that it invalidates the credibility of having different event severity. For example, if a datasource of a SNMP check reaches a warning state (the value of the datasource would equal 1), then a warning event will be created, which is ok. On the other hand if a datasource is in an unknown state (the value of the datasources would equal 3) then three events get created (one for each threshold that has been exceeded).
Has anyone has any issues with this?
>
Just a clarification for Nagios exit codes: http://dev.zenoss.org/trac/ticket/5294 (this is in docs for the upcoming 2.5 release)
And in addition to the Point Threshold ZenPack there are the excellent Range Threshold and Predictive Threshold ZenPacks.
Follow Us On Twitter »
|
Latest from the Zenoss Blog » | Community | Products | Services Resources | Customers Partners | About Us | ||
Copyright © 2005-2011 Zenoss, Inc.
|
||||||||