- To monitor a sysplex, start with the Sysplex Summary (SYSSUM)
report. Use the report options to set the Performance Index threshold
to a value of, for example, 0.8 as a warning level, and select a
type, so that service class periods are included in the report.
Start the report in GO mode and let it run.
As long as everything is running well, the performance status line at the top of the report shows only green. When the "warning" level for a goal is reached, the corresponding service class with the respective period appears on the report in yellow. And when a goal is not met, the corresponding service class appears on the report in red, followed by the service class period that missed the goal.
- To find out what is causing the red line, leave GO mode and put
the cursor on the line where the goal was not met. If several goals
have been missed, the performance index can be of help to find out
which goal was "missed most".
Depending on the type of service class, different detailed reports are shown:
- For service classes, a response time breakdown is shown on the Response Time Distribution (SYSRTD) report.
- For subsystem service classes, the transaction states are shown on the Work Manager Delays (SYSWKM) report.
The SYSRTD report has a sysplex view in the upper part of the screen, and provides a single-system breakdown in a scrollable list on the bottom part of the screen.
Furthermore, you can step from the SYSWKM report to the SYSRTD report using cursor-sensitive control, if you need some information from that report for additional investigations, or you want to continue navigation from that report.
- The scrollable section in the SYSRTD report is the link from the sysplex level to the single system. From these lines it is possible to "zoom" into any of the listed systems. Placing the cursor on the system-ID in one of the rows and pressing ENTER, leads to the SYSINFO report of that system, thus allowing further analysis based on the data shown there. Placing the cursor on a specific data column in one of the rows of the scrollable area leads to a specific report of that system that provides additional information related to the selected column.
- Finally, when the single-system level is reached, navigation among those reports is possible as described in step 5.
- In the workflow/Exceptions (WFEX) report, you can
identify jobs and resources with low workflow values or jobs that
have met exceptional conditions. For example, you can check the Reason
field to identify the user or the possible cause of delay. Once you
recognize a user or a resource with a potential problem, you can analyze
the situation using cursor-sensitive control. Using cursor-sensitive control describes
how to invoke reports using this method.
If you are on the Delay report, check the delay value (for PROC, STOR, DEV, SUBS, OPER, ENQ) with the largest value associated with a job, use cursor-sensitive control to navigate to the Job Delay report for that type of delay to analyze the main reason for the delay.
In case of a delay due to devices (DEV) or enqueued resources (ENQ), you can use cursor-sensitive control to further investigate a problem by looking at the resource-oriented device report (DEVR) and the resource-oriented enqueue report (ENQR). For storage problems involving paging or swapping delays, you can use the resource-oriented storage delay report (STORR). Use either the job entry subsystem (JES), hierarchical storage manager (HSM), or cross-system coupling facility (XCF) delays report for a delay associated with SUBS. For OPER delay, use cursor-sensitive control to see the appropriate Job Delay report.
Figure 1. Suggested Sequence for Using Monitor III Reports
For a summary of common system storage consumption, use the Common Storage Summary report (STORC). To identify remaining storage, use the STORCR report.
For a summary of how the workflow and delay of the measured system affects performance, use the SYSINFO report.
Figure 1shows a suggested sequence for using Monitor III reports to resolve potential problems.