Performance and Diagnostic Advisor
Overview
The Performance and Diagnostic Advisor runs in the JVM of the appserver and periodically
- Checks for inefficient settings
- Issues recommendations as standard product warning messages
These recommendations are displayed...
- As warnings in the admin console under Runtime Messages in the WAS Status panel
- As text in the appserver SystemOut.log file
Enable the Performance and Diagnostic Advisor has minimal system performance impact.
We can use the Performance and Diagnostic Advisor to enable the lightweight memory leak detection.
Contradictory advice
Performance and Diagnostic Advisor gives advice on a server scope. In a WAS ND environment we might receive contradictory advice on resources that are declared at the node or cell level, but used at the server level. For example, consider a data source declared at the node level that has a connection pool size of {10,50} and is used by two servers...
- server1
- server2
If server1 uses only two connections and server2 uses all fifty connections during peak load, the optimal connection pool size is different for the two servers. Therefore, the Performance and Diagnostic Advisor gives two sets of advice (one for server1 and another for server2). The data source is declared at the node level and make the decisions appropriately by setting one size that works for both, or by declaring two different data sources for each server with the appropriate level.
Subtopics
Diagnostic alerts
Enable Performance and Diagnostic Advisor
Related concepts
Performance advisor types and purposes
Lightweight memory leak detection