effective load average monitor
I'm working on trying to get an effective Load Average monitor working in OpManager. I did email support but the answer they sent back indicated they didn't really read my questions.
Here are the relevant entries from my conf files. My command line works. If I run this command line on a Linux hosts, it gives me back correct looking output. If I run this through OpManager, I get unexpected output and nothing gets graphed.
The gist of this is I'm trying to come up with a metric based on Load Average (15 minutes) and number of CPU's to come up with a load percentage. Yes, this can exceed 100% which should be an effective threshold.
Here are the relevant entries from my conf files. My command line works. If I run this command line on a Linux hosts, it gives me back correct looking output. If I run this through OpManager, I get unexpected output and nothing gets graphed.
The gist of this is I'm trying to come up with a metric based on Load Average (15 minutes) and number of CPU's to come up with a load percentage. Yes, this can exceed 100% which should be an effective threshold.
<CATEGORY ID="System Load" command="echo `grep processor /proc/cpuinfo | wc -l` `/bin/cat /proc/loadavg | awk '{print $3}' | sed 's,\.,,'`" parseAll="true" description="CPU load due to the jobs handled periodically by the CPU.">
<ENTITY ID="LoadAvg" applyThresholdOver="(curValue/cpucount)" threshold="Th6" storeHistory="true" storeValue="(curValue/cpucount)" units="%" comparePreValue="false">
<PARSE token="1" name="cpucount"/>
<PARSE token="2" name="curValue"/>
</ENTITY>
</CATEGORY>
<DEFAULTGRAPH Name="Telnet-LoadAvg">
<YAXISTEXT>%</YAXISTEXT>
<DISPLAYNAME>System Load</DISPLAYNAME>
<name>LoadAvg</name>
<Vendor>Telnet</Vendor>
<Monitor>Resource</Monitor>
<Description>CPU load due to the jobs handled periodically by the CPU.</Description>
<oid>System Load</oid>
<Provider>CLI</Provider>
<type>node</type>
<interval>900</interval>
<THRESHOLDENABLED>false</THRESHOLDENABLED>
</DEFAULTGRAPH>