We are running on AC classic V7 on CentOS 7. Recently we upgraded from the Gold Standard Release to 9346 release.
After the upgrade, all the business logic seems to working as expected but we noticed a huge CPU usage jump on the application server. Checking the issue a bit deeper, we find that it is the “web” process on the application server consuming most of the CPUs. The web process runs a couple of "http-nio-8080-e” threads and each time a SOAP request comes in from the AC client console, there seems to be very heavy XML related work going on that can briefly saturate 1 vCPU on an AWS EC2 machine (e.g. 90% usage according to top). When we have multiple AC users working concurrently with their client console, this high CPU usage issue can easily saturate a VM with 8 vCPUs.
Are there other users facing similar issue with this relatively latest relase?
Looking internally, I have not seen any other client's report any issues with the latest release regarding CPU spikes. I would recommend opening a ticket with Campaign Support with exactly what you are seeing, any server info, the prior build version, and if possible any performance benchmarks prior to the upgrade.
Thanks a lot for the response. We opened a ticket and let's see if we can get some hints on how to investigate this issue further.
I got more feedback from our AC operators. So it seems the high CPU usage scenarios is more easily triggered if they use the "dashboard" UI of the client console, not so much issue if they use the "explorer" UI of the client console.
Is there any documentation telling the difference between these two different views? Is it possible that some complications were introduced in the "dashboard" UI (or the way requests were handled in the web process) in the 9346 build?