We have a weblogic cluster across 6 boxes. Each box has 3 JVMs. 2 JVMs on box 4 are showing very high CPU utilization (in the range of 80-90%) compared to other JVMs which have less than 10% CPU utilization.
We checked load-balancing. The requests are getting uniformly distributed to all the JVMs. All the JVMs are doing GC correctly, there is no issue with Garbage collection. Each JVM has the same GC and memory configuration.
Is there any way to figure out what threads are utilizing high CPU? We can not restart the JVMs or modify any settings on them as those are the production JVMs.