|
|
| Subject: | Configure a threshold on a DataPoint Alias? or other transform? |
| Author: | [Not Specified] |
| Posted: | 2014-09-29 18:25 |
Here's the issue...
I have several linux hosts with core counts ranging from 2 to 16
The current OID used to pull load is showing the accurate 5 min load average. Which can range all over the map for these hosts. The issue here comes down to the threshold to alarm for high load. a load of 2 on one of the 2 core systems should be alerted, but a load of 2 on the 16 core systems should not.
I am looking for a way to make the threshold relate to the number of cpu's on the host.
A Data Point alias looks like the right place for such relative math transform: ${here/hw/Processors},/
But I can't see how to make a DataPointAlias part of a Threshold.
Ideas
| Subject: | I'll second this one |
| Author: | [Not Specified] |
| Posted: | 2014-10-01 12:14 |
I would like help with the exact same question. Although I gave up on using SNMP to monitor linux nodes. For one the default template uses a deprecated OID for CPU Idle percentage data, but regardless of that the SSH linux load threshold has the same shortcoming that you described for the SNMP monitor.
| Subject: | http://wiki.zenoss.org/Linux |
| Author: | Jay Stanley |
| Posted: | 2014-11-25 12:34 |
| < |
Previous Zenoss 4.2.5 SP203 Add Multiple Devices Broken |
Next zenrrdcached not running |
> |