The dominating consumers so far weren't contained in the application or JVM subsystems. If the CPU time is not overly spent in the kernel or idling, we start investigations in the JVM.

Since the JVM is responsible for memory management, its performance will indicate potential memory issues. Mainly Garbage Collection (GC) logs, together with JMX tooling help investigate scenarios.

Memory leaks will lead to increasing memory usage and excessive garbage collector runs that occupy the CPU. Inefficient memory usage will equally lead to excessive garbage collections. The GC executions ultimately cause the JVM being the dominating consumer of the CPU.

This is another example of why it's important to follow the process of the jPDM decision tree. The performance issue arises in high CPU usage, although the actual bottleneck in this case is the memory leak.

As of today, the main cause of performance issues are related to memory, mostly from application logging that results in extensive string object creation.