Researching a performance problem

Most of the tools that collect or analyze performance use either trace or sample data. Collection Services regularly collects sample data on a variety of system resources. Several tools analyze or report on this sample data, and you can use this to get a broader view of system resource utilization and to answer many common performance questions. IBM® i Job Watcher and IBM i Disk Watcher also collect sample data. For more detailed performance information, several tools generate trace-level data. Often, trace-level data can provide detailed information about the behavior and resource consumption of jobs and applications on your system. Performance Explorer and the Start Performance Trace (STRPFRTRC) command are two common tools for generating trace data.

For example, if your system is running slowly, you might use the IBM Navigator for i monitors to look for problems. If you see that the CPU utilization is high, you could identify any jobs that seem to be using an unusually large amount of resources. Then, you may be able to correct the problem by making configuration changes. However, some problems will require additional information. To get detailed information about that job's performance you could start an IBM i Job Watcher collection for the desired job, gather detailed information about that job's behavior on the server, and potentially make changes to the originating program.