I have one more small query.
If I use this expression to do my alerting, the value that I will get when
I use $value in my alert will be Load per CPU Core.
How to get the actual CPU Load value itself while using the expression
mentioned by you.
On Wednesday, March 11, 2020 at 11:07:21 PM UTC+5:30, Yagyansh S. Kumar
wrote:
>
> Maybe I'll refine the threshold even further but for now this works.
> Thanks a lot for help.
>
> On Wednesday, March 11, 2020 at 10:47:14 PM UTC+5:30, Harald Koch wrote:
>>
>>
>>
>> On Wed, Mar 11, 2020, at 13:00, Yagyansh S. Kumar wrote:
>>
>> Hi. I have configured alert for CPU Load for my servers and my current
>> threshold is 8 for warning and 10 for critical.
>> I want to make this threshold dynamic i.e I want the critical alert when
>> the CPU Load becomes greater than the number of CPU Cores of the machine.
>> Eg. For a server with 8 CPU cores, I want a critical alert when CPU load
>> > 8 and for a machine with 16 CPU cores, I want a critical alert when CPU
>> Load > 16.
>>
>>
>> count without (cpu, mode) (node_cpu_seconds_total{mode="system"}))
>>
>> is one way of counting the number of CPUs on a server, so:
>>
>> (node_load5 / count without (cpu, mode)
>> (node_cpu_seconds_total{mode="system"})) > 1
>>
>> would alert you when the (5 minute) load average is greater than the
>> number of CPUs.
>>
>> I find this alert to be noisy - your experience may differ.
>>
>> --
>> Harald
>>
>>
>>
--
You received this message because you are subscribed to the Google Groups
"Prometheus Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/prometheus-users/10f18943-1285-4354-9f90-d67d92cca7a9%40googlegroups.com.
