DX Application Performance Management

In the Investigator Metric Browser, how is the "Agents Stats|Resources" metric populated with data?

Is this a field customizable? These also appear to be tied to the "Resources" tab. I thought the EM file "ResourceMetricMap.properties" populated this Tab, but it appears to not be the case. Any insight or suggestions?

Hi Scott,

We have the GC Monitor installed on our servers, so I get the metric for "% Time Spent in GC" from the GC Monitor.

Also CPU Utilization% would usually show up under the node CPU. See Screen Shots below.

Are you able to see these metrics too, along with the ones that show up under Agent Stats?

Thanks,

KIRAN T

Also to add to that info, All I see under Agent Stats is "Build and Release".

Thanks,

KIRAN T

Yes, the other metrics are visible in the other, usual places.

Scott,

Your understanding is correct. As of APM version 9.1 and higher, there are now up to four new metrics under [font=Courier New]Agent Stats|Resources[font] which also populate the [font=Courier New]Resources[font] tab when you have an agent selected in the [font=Courier New]Metric Browser[font]. The four metrics are below:

[font=Courier New]Agent Stats|Resources:% CPU Utilization (Host)
Agent Stats|Resources:% Time Spent in GC
Agent Stats|Resources:Threads in Use
Agent Stats|Resources:JDBC Connections in Use[font]
These metrics are, in effect, like JavaScript calculated metrics - they are derived from other existing metrics and get reported to the same agent. The logic used to determine which existing metrics get used is defined in the [font=Courier New]ResourceMetricMap.properties[font] file. By default, the logic is as below:

[color=#008000]#CPU Utilization[color]

cpu.utilization.path.1=CPU:Utilization % (process)

[color=#008000]#Time Spent in GC[color]
memory.utilization.path.1=GC Monitor:Percentage of Time Spent in GC during last 15 minutes

[color=#008000]#Threads in Use[color]
threads.used.path.1=WebSpherePMI|threadPoolModule|WebContainer:ActiveCount
threads.used.path.2=WebLogic|JMX Aggregate|Thread Pool:Waiting Request Count
threads.used.path.3=Tomcat|ThreadPool|default thread pool:getCurrentThreadsBusy

[color=#008000]#JDBC Connections in Use[color]
connections.used.path.1=WebSpherePMI|connectionPoolModule:WaitingThreadCount
connections.used.path.2=WebLogic|JMX Aggregate|JDBC Connection Pool:Waiting Thread Count
If the specified metrics do not exist on the agent for any of the four resource metric types, the respective resource metric will not get generated. For example, if you had [font=Courier New]GC Monitor[font] disabled, you would not get the [font=Courier New]% Time Spent in GC[font] metric. Similarly, if your application server is not WebSphere, WebLogic, or Tomcat, you will not get the [font=Courier New]Threads in Use[font] metric. However, you can add your own source metrics here, which can be very handy. Since many of the application I work with use Tomcat but do not use Tomcat's JDBC Connection pooler, I haved use a JavaScript calculator to sum up the number of JDBC backend connections to produce a [font=Courier New]JDBC:Connection Count[font] metric and then used that metric to tie it to the [font=Courier New]JDBC Connections in Use[font] metric. To my knowledge, wildcards are not supported in this configuration file, which is unfortunate and would have made that process much simpler.

On a related note, the Agent Stats node is also where the Sustainability metrics will appear, if you ever have to turn those on.

Thanks jakbutler.

I thought this was how it should work, but I was thrown when I commented out CPU in ResourceMetricMap.properties and it still showed up in the dashboard.

#cpu.utilization.path.1=CPU:Utilization % (process)

I've got a sandbox system set up that'll allow me to zero in on the right source metrics.

I would mark this "Resolved" :bashful:

Scott,

What I don't remember is whether or not you need to restart the Enterprise Manager for changes to that file to take effect. I seem to recall having to, but it's been a while.

Since you're the original poster, feel free to mark one of the above posts as the resolution, and let us know if you have more questions or discover something else on this topic.

- Jack

That is absolutely true. The reason why I dont see these metrics under Agent Stats is that I am on Version 9.0.7.2.

Thanks,

KIRAN T

i would like to know how the CPU Utilization % process is getted ... it isnt the same as the one that is getted from top.
so i mean the top says that the process is consuming 2% of cpu
and the graphic says it is always 0 .... the behaviour is worse with more load .....

Oscar,

You'll receive the answers to your question in your original thread: https://communities.ca.com/web/ca-wily-global-user-community/message-board/-/message_boards/view_message/100435183