• Instance level monitoring – performance captured for individual user transactions with no aggregation or averaging.
  • Trace problem transactions across JVMs and remote hosts – follow problem transactions across application tiers to diagnose complex problems in a distributed environment.
  • Understand context around slow methods, like slow SQL.
  • Find sporadic problems that are hidden by averaging transactions.
In addition, the Deep-Dive data includes metric data to give insight into the overall health of your system and help explain performance bottlenecks:
  • JVM heap sizes and garbage collection (GC) event durations over time.
  • Many other performance-related JMX metrics such as JMS message queues, EJB pools and DataSource Connection pools.
Below is an example of the transaction tracing data that might be associated with the Call Summary Profile above:
Deep-dive perspective, transactions sorted by Elapsed Time
Deep-dive tracing can be essential to identifying the root cause of a performance problem in an enterprise application, but is best applied in a targeted way to avoid the risk of further degrading application performance by collecting unnecessarily detailed performance data.

Note

Even if you haven't enabled tracing, you can still use the Deep Dive perspective to examine garbage collection (GC) and JMX metrics.


loading table of contents...