CPU 100% after server running 2 weeks

Support
#1

I’m using v6.8.1

I know the likely first step is to upgrade to the latest 6.8.x release (not yet ready to migrate to 6.9/6.10), but wanted to raise a question to see if anyone has seen behaviour such as below:

I have a server running - web and middleware layer on the same machine and DB on a separate server. The web/middleware server is running under Docker.

The server can run for around 2 weeks, then the CPU, which sits around 5-10% usually, suddenly spikes to 100% and the server effectively is unusable and needs a restart (docker stop and docker run) to recover. There is nothing obvious in the logs as to what is causing the spike and the usage is normal user usage on a daily basis.

Has anyone seen a server running for a period of weeks causing this problem, or any way to try to diagnose?

Many thanks

#2

Hi,

We have a lot of long-running systems based on CUBA platform in production. And from our experience, there is no common reason for performance problems, it is highly dependent on your workload.

I’d recommend that you attach VisualVM to your sever and try to perform Sampling.

#3

Also,

Please note, that all default platform settings are for development. Usually, you should tune your application, web server and DB in production environment.