PMI architecture

The Performance Monitoring Infrastructure (PMI) uses a client-server architecture.

The server collects performance data from various WebSphere® Application Server components. A client retrieves performance data from one or more servers and processes the data. WebSphere Application Server, Version 6 supports the Java™ Platform, Enterprise Edition (Java EE) Management Reference Implementation (JSR-77).

In WebSphere Application Server, Version 6 and later, PMI counters are enabled, based on a monitoring or instrumentation level. The levels are None, Basic, Extended, All, and Custom. These levels are specified in the PMI module XML file. Enabling the module at a given level includes all the counters at the given level plus counters from levels following the given level. So, enabling the module at the extended level enables all the counters at that level plus all the basic level counters as well.

JSR-077 defines a set of statistics for Java EE components as part of the StatisticProvider interface. The PMI monitoring level of Basic includes all of the JSR-077 specified statistics. PMI is set to monitor at a Basic level by default.

[AIX Solaris HP-UX Linux Windows]As shown in the following figure, the server collects PMI data in memory. This data consists of counters such as servlet response time and data connection pool usage. The data points are then retrieved using a web client, a Java client, or a Java Management Extensions (JMX) client. WebSphere Application Server contains Tivoli® Performance Viewer, a Java client which displays and monitors performance data. See the Monitoring performance with Tivoli Performance Viewer, Third-party performance monitoring and management solutions, and Developing your own monitoring applications topics for more information on monitoring tools.
PMI architecture
The figure shows the overall PMI architecture. In one section of the picture, the server updates and keeps PMI data in memory. The other section displays a web client, a Java client, and a JMX client retrieving the performance data.