DX Application Performance Management

I would like to better understand how the investigation of memory leak problem works under the Introscope. In the chart below we have a memory pool view:

Based on the graph we have a first (1) vision for research in the Health Check Introscope. We can say that a memory leak?

What would be a recommended next step for this analysis

Oi Ricardo:

Looking at wikipedia's definition Memory leak - Wikipedia, the free encyclopedia , I would expect to see more memory being retained, not released until none is available.

Many metrics go up when there is a memory leak

Average Response Time, Bytes in Use, and CPU Utilization are high Instance Counts are increasing (until GC) Responses per Interval are low. In some cases there should be a steady increase.

What do you see for the metrics that I mentioned?

Hi,

You're looking at the survivor spaces which are small temporary spaces that should produce always constant metrics (more or less).

It's important to fully understand how the JVM handle memories, you should have a look to the following site

Java Garbage Collection Basics

Simply put, a Java memory leak pattern is when the old generation space keep growing after each "old gc" while we know the heap size should be enough for the application.

Without entering at the generation levels, from the high level point of view, the % of Java Heap space has less and less memory after each GC.

Hi Ricardo:

   Two responses have been given and there has been no follow-up questions from you. So marking as answered...