CPU Load avg. monitoring CDM Probe

[Sid]

I am not able to find anything related to inodes and Load average monitoring for cdm probe in techdocs.
Can someone please help me understand how exactly both work. the commandsets and other things that probe uses .
Need to understand about Load average thresholds also

------------------------------
HCL Technologies
------------------------------

[Stephen Danseglio]

Hi Sid,

Do you need more details than what is listed here?

https://techdocs.broadcom.com/content/broadcom/techdocs/us/en/ca-enterprise-software/it-operations-management/ca-unified-infrastructure-management-probes/GA/alphabetical-probe-articles/cdm-cpu-disk-memory-performance-monitoring/cdm-ac-configuration.html

Steve

------------------------------
Support Engineer
Broadcom
US
------------------------------

[David Michel]

also there is this:
CPU, Disk and Memory probe (cdm) Technical Brief
Article Id: 35051
https://ca-broadcom.wolkenservicedesk.com/external/article?articleId=35051

------------------------------
Support Engineer
Broadcom
------------------------------

[Sid]

Hi David,
I wanted to understand how CPU load average is monitored ? how does it calculate cores ?
and what does the .10/.20 etc of CPU load average mean?

------------------------------
HCL Technologies
------------------------------

[Gregg Stillwell]

HI Sid,

I have always defaulted to the vendor or the OS. Here is the doc which tells us what command cdm runs to get the OS's value for load average. To my knowledge UIM does not do a "backroom" calculation to figure this out. I would use the "man" command to list each command that cdm uses to get the "load" on attached doc. Or google the commands. I have forwarded these outputs to supply the explanations, for the questions you are asking.

https://knowledge.broadcom.com/external/article?articleId=11858

[Garin Walsh]

The document referenced above is also consistent with my experience. 

And generally speaking, load average is the average number of processes ready to run at any point in time. Without knowing anything more about the system, this number is completely useless except to compare to what it was in the past and even then you don't know if that comparison indicates a problem or not.

As an example, if your load average is 4, can you say if that's good or bad? No, because it depends on the number of CPU cores available - if your server has 16 cores, a value of 4 means your CPU usage is only 25% maximum. If your server has 2 cores it means that the server is being asked to do twice the work it is capable of at that point in time. Similarly if you load average was 4 and changed to 8 is that bad? No idea because if you had the same 16 core server, you are now only at 50% utilization. On the other hand if your server has only 2 cores, you're probably in a bad situation.

Looked at over a long period of time (days to weeks) you might be able to forecast future resource demands but that would depend on having a consistent demand on resources that's predictable and most server workloads are not.

[Sid]

Than you Garin for sharing the detailed knowledge. :)

------------------------------
HCL Technologies
------------------------------

[Sid]

Thanks a lot Gregg, its understood :)
Thanks again.

------------------------------
HCL Technologies
------------------------------

[Guillaume Briere]

Hello,

You can see the load average value by running "top" at the command line. It's at the end of the first line.

You can read this document that explain what is the load value: https://scoutapm.com/blog/understanding-load-averages