I´m looking for any way to use Orion to create a Solaris/Red Hat CPU´s process run queue
graphics.
Thanks
I don't follow you. What are you trying to do?
http://www.metron-athene.com/reference/performance_tips/unix/unix_tip07.html
Hello Contracer,
the Net-SNMP MIB propose the MIB table: Loadaveload - 1.3.6.1.4.1.2021.7.3
it is supposed to return "The 1 , 5 and 10 minute load averages." If that table is supported simply create a UnDP to poll the relevant values.
I found this from the following article: http://www.net-snmp.org/wiki/index.php/FAQ:Agent_32
If the table is not supported....
A solution would be to use the Extend feature in Net-SNMP which calls a script that returns you the CPU Run Queue using the uptime command.
Net-SNMP Extend Feature description:
http://www.net-snmp.org/docs/man/snmpd.conf.html scroll down to "Extending Agent Functionality"
Script example using awk to parse the output:
http://bash.cyberciti.biz/monitoring/get-system-information-in-html-format/
You can then retrieve the values using a Universal Device Poller and start graphing the results.
UnDP Tutorial: http://www.solarwinds.com/resources/democenter/Orion/UniversalDevicePollerIntro/index.html
HTH,
Yann
Okay! I have up and running run queue UnDP poller and alerts works! great! Now question if i have more CPU on box then this number should change right ? for example if i set alert run queue greater than 20 and if i have 8 CPU in my box so in ratio of CPU my run queue is very low for this box. If i have 1 CPU then this condition is correct.
Is there anyway we can find out CPU and make alert base on CPU ratio?
