Hello, I am trying to make sense of this section in APM. I am monitoring some processes that say they use up 98% of CPU, however on the monitored server the CPU load is very low.
Is the application running within a certain segment of the CPU, and this is telling me how much of that segment is being used?
Any information somebody may have would be great. Thanks.
What monitor are you using: WMI Process Monitor, WMI Service Monitor, or SNMP Process Monitor?
What service pack do you have installed?
I am using the SNMP Process monitor on a Linux server. The APM is on the latest version with the service pack installed.
rkawell:on the monitored server the CPU load is very low
Does the node details page show the CPU load for the server as low or were you looking at the CPU load on the actual machine?
The CPU numbers we report are what the SNMP agent is telling us. Is the process constantly at 98%?
As for real CPU load on the server, I looked at both the server and the node details in Orion. It is just the CPU process in the application monitor looking at one process. For example, OpenVPN fluxuates based on time of day, and how busy the network is, but most of the time the OpenVPN application is running at 100% of the CPU that is reported in the application monitor, not on the server itself.
Are there multiple instances of this process running?
Are you using Net-SNMP? What version?
Does this happen with all processes you try to monitor on that box or just the OpenVPN process?
I noticed the same thing. I am running the latest version of net-snmp on RHEL4 server. The multiple oracle processes are reading the CPU load of each process well. However, any and all java processes on all java application servers reflect the process CPU load is greater than 90% all the time. The NPM reports of the server's overall CPU load is less than 10% on each of these servers. Normally there are 1-3 processes running java per server.
This is not new in APM because it was the same in the legacy Application Module 8.5.
I guess I'm just looking for an explaination of what it is representing. If my process is running out of perceived CPU space then couldn't that be a problem? I would think at some point we would need to adjust the application to think it can use more CPU threads. Thoughts?
Did you get to the bottom of this? I am running into the same issue monitoring CPU load on RedHat servers using SNMP. Same issue in both AM and APM.
I too am experiencing this problem.
I am running net-snmp to monitor processes on a linux box. Right now it is reporting that one of my processes is using at 93% of th CPU, but when I use the SNMP Real-time graph with the OID HOST-RESOURCES-MIB:hrSWRunPerfCPU.<pid>, I get the correct results.
I'm using Orion 9, SP2.