I have a task to monitor for excessive processor queue's on Windows 2003 servers.
The request is stated as below.
For a group of servers
If the processes in waiting count is more then two times the number of cores for over two minutes alert via email.
seems simple enough.
I'm working with Perf_counter due to some of the necessary default zabbix counters not working at all in a Windows environment.
{s-vcenter
erf_counter[\System\Processor Queue Length].avg(60)}>60 / {s-vcenter:system.cpu.num[online].last(0)}>2
I started with the above in the trigger expression to get things going.
It seems the math isn't calculated at all, though the manual alludes to this being possible and several forum posts as well.
Is my expression just plain wrong? What would the corrected expression look like?
Is there a better way to attack this trigger?
Zabbix 1.6.2
The request is stated as below.
For a group of servers
If the processes in waiting count is more then two times the number of cores for over two minutes alert via email.
seems simple enough.
I'm working with Perf_counter due to some of the necessary default zabbix counters not working at all in a Windows environment.
{s-vcenter
erf_counter[\System\Processor Queue Length].avg(60)}>60 / {s-vcenter:system.cpu.num[online].last(0)}>2I started with the above in the trigger expression to get things going.
It seems the math isn't calculated at all, though the manual alludes to this being possible and several forum posts as well.
Is my expression just plain wrong? What would the corrected expression look like?
Is there a better way to attack this trigger?
Zabbix 1.6.2
Attached Files