In one of the database servers "_ CPU" was monitored through logmon Probe

Document ID : KB000128826
Last Modified Date : 15/03/2019
Show Technical Document Details
Issue:
Customer needs to monitor:

- CPU Usage Total/Aggregate / Individual 
- CPU load average, 1, 5 and 15 minutes 
- System Load 
- Uptime 
Environment:
- UIM 8.5.1 or higher
- cdm 6.30 or higher
Cause:
- Not clearly documented, yet all of these monitors/metrics are supported by the cdm probe
Resolution:
The cdm probe allows you to monitor: 

- CPU Usage Total/Aggregate or Individual 
- CPU load average, 1, 5 and 15 minutes 
- System Load 
- Uptime 

Please refer to: 

https://docops.ca.com/ca-unified-infrastructure-management-probes/ga/en/alphabetical-probe-articles/cdm-cpu-disk-memory-performance-monitoring/cdm-ac-configuration 

....(Linux, Solaris, AIX and HP-UX platforms)
Select the required monitor and select Publish Data and Publish Alarms fields to retrieve QoS data and generate alarm messages, respectively.
cdm probe:

CPU Usage (out of the box, but you must publish the QOS data)

CPU Load Average
  For Linux, Solaris, AIX and HP-UX platforms, you can configure the following CPU load average monitoring parameters:

Load Average
- Max. Load Average: specifies the maximum average load on the CPU in the monitored duration, options include 1, 5 and 15 minutes.

- CPU Load average per processor

 System Load
- Max. System Load: specifies the maximum load on the CPU

Note: The load average monitors for 1, 5 and 15 minutes become available in the cdm GUI ONLY if the 'Calculate Load Average per Processor' checkbox is selected in the cdm probe General Tab in IM. Also, Via Raw Configure, the option is available/listed under the <setup> section as 'load_average_per_processor.' It can be set to yes to enable it.

In the cdm probe via Raw Configure, under the <cpu> section:

cpuTotalQoSTargetSet
- Must be set to 'yes' to expose the Load Average options in the Infrastructure Manager (IM) GUI.

In the cdm GUI, under the Advanced Tab, you can enable CPU QOS collection and alarm thresholds for load averages of 1, 5 and 15 minutes.

Via Raw Configure, under the <cpu> section, the load average options can be enabled for 1, 5 and 15 minutes by setting them to yes.

qos_load_average_1min
- Load Average 1 min: monitors the average CPU usage during the last one minute.
QOS: QOS_LOAD_AVERAGE_1MIN 
Unit: Count 
Descr: The average system load over the last one minute. 

qos_load_average_5min
- Load Average 5 min: monitors the average CPU usage during the last five minutes.
QOS: QOS_LOAD_AVERAGE_5MIN 
Unit: Count 
Descr: The average system load over the last five minutes. 

qos_load_average_15min
- Load Average 15 min: monitors the average CPU usage during the last fifteen minutes.
QOS: QOS_LOAD_AVERAGE_15MIN 
Unit: Count 
Descr: The average system load over the last fifteen minutes. 

And, via Raw configure yiou can see the load average options/parameters:

qos_load_average_1min 
qos_load_average_5min 
qos_load_average_15min 


Note that as of UIM 9.0.2, these options are available in the IM via Raw Configure, as well as visible in the GUI when they are enabled but they will not display/show up in the Admin Console.

What are the commands used by the cdm probe to calculate these metrics? (Linux, Solaris, AIX, and HP-UX)

The CPU usage information is found by reading the file “/proc/stat”. Based on the data gathering interval either the 1, 5 or 15 minute average is used.

The load data is found by reading the file “/proc/loadavg”. Based on the data gathering interval either the 1, 5 or 15 minute average is used.
Additional Information:
To understand which cdm monitoring options/metrics are supported on a given platform go to:
 
https://docops.ca.com/ca-unified-infrastructure-management-probes/ga/en/alphabetical-probe-articles/cdm-cpu-disk-memory-performance-monitoring/cdm-metrics