Dawn May | Published March 20, 2019
The Performance Data Investigator is a browser-based graphical user interface for viewing and analyzing IBM i performance data. The Performance Data Investigator is a general name for the function; Investigate Data is the actual task name.
The Performance Data Investigator (PDI) is included with the IBM Navigator for i web console, which is part of the IBM i operating system. Everything you need to use this web console is installed by default (see “IBM Navigator for i” in the Resources section). The HTTP admin server is required, and it is automatically started.
Simply point your browser at your IBM i partition, using the 2001 port (for example, http://<mysystem>:2001) where mysystem is the IBM i partition host name. You get a sign-on screen where you enter your IBM i user ID and password. After signing in, you can see the Welcome panel and the tasks available in the left navigation pane.
Figure 1. IBM Navigator for i Welcome page
In the left navigation pane, you can find the Performance tasks. Click the box to the left of Performance to expand the section and you will find the Investigate Data link. Click Investigate Data to open PDI in the main Navigator page, or you can expand the Investigate Data task and work with the PDI from the left navigation pane, as shown in Figure 3.
Figure 2. Performance tasks
Figure 3. Investigate Data tasks
IBM i has four different kinds of performance data that you can analyze with PDI. The ability to collect this performance data is provided with the operating system. The Performance data collectors are:
The Performance Data Investigator has different features available depending upon the Performance Tools options you have installed on your IBM i partition. The graphical user interface for Collection Services is included with the base operating system. However, the GUI for the other performance data collectors requires the appropriate option of the IBM Performance Tools Licensed Program Product to be installed.
The following are the Performance Tools options required to use PDI with various kinds of IBM i performance data:
PDI has additional capabilities, such as working with Database performance data and the Batch Model tool. You must have the IBM Performance Tools Licensed Program Product, Manager feature, installed for these to be available.
In addition to having the required options of the IBM Performance Tools Licensed Program Product installed, you also need to ensure that users of PDI without a powerful user profile (for example, *ALLOBJ users) are added to the QPMCCDATA authorization list. Use EDTAUTL QPMCCDATA to edit the authorization list to add users.
The IBM i wait accounting technology can help in identifying reasons for issues (such as a record lock, object lock, disk I/O operation, journal operation, and so on) that result in a thread or task not using the processor. Because threads and tasks can wait for a wide variety of reasons, wait technology can be very useful in understanding the wait conditions and possibly eliminating or reducing wait time. Eliminating wait time can improve performance.
Within the PDI, there are many charts that display wait information. Having a basic understanding of wait accounting helps you better understand those charts. See the Resources section for additional information.
When you start the Investigate Data task, you will see a list of perspectives. A perspective is a chart or table of the performance data that is displayed. Perspectives are grouped into folders, and a collection of perspectives is called a content package. Each performance data collector has a content package. The content packages you see when you start the Investigate Data task depends upon the options (if any) of the IBM i Performance Tools licensed program product that you have installed (see the prerequisites discussed earlier). Figure 4 shows all of the possible content packages that are available.
Figure 4. Investigate Data content packages
For the purposes of this article, the discussion centers around the Collection Services content package, because that is included with the operating system and does not require the Performance Tools licensed program product (LPP).
To see the list of perspectives in a content package, you must click the folder icon in front of the content package name. It then expands to show a list of nested folders where perspectives are grouped by the type of performance data they display. Figure 5 displays the list of perspectives in the Collection Services content package. Within the Collection Services content package, the perspectives are organized by the type of data they display.
Figure 5. Collection Services perspectives
At the bottom of the Investigate Data panel is where you can specify the library in which the performance data resides and the name of the collection you want to investigate. The default library is QPFRDATA and the default collection is the most recent one, which allows you to look at the most recently collected performance data.
The button on the main Investigate Data panel allows you to display the configuration options for the behavior of PDI. Most of the options are self-explanatory. However, Enable Design Mode deserves a brief mention. Design mode allows you to create your own content package and perspectives. The details of the design mode are far beyond the scope of this article.
The Search button on the main Investigate Data8 panel allows you to access the metric finder to search for performance metrics and identify the charts, tables, or files in which the metric exists. While the metric finder remains available, the Investigate Data Search task is a superior search function. Investigate Data Search provides an interface where you can search on terms within all areas of PDI, providing a much more comprehensive search feature.
Figure 6. Investigate Data Search
Navigator for i has the ability to save favorites to make it easier to find the tasks you use most often. You can also save PDI perspectives as favorites. You do this from the main PDI window by clicking the Save as Favorite button.
As noted earlier, you can also navigate to a perspective with Investigate Data in the left pane. If you navigate to various perspectives in the left pane, you are required to select the collection for each perspective you display. If you navigate to various perspectives in the main window, PDI remembers the collection you selected, and so you can move from chart to chart without changing the collection.
Displaying performance data with PDI has many features for interacting with the displayed data, switching between chart and table views, data customization, and more. This section briefly reviews the main features of PDI when displaying performance data.
A good starting point with Collection Services data is the CPU utilization and Waits Overview perspective. This chart shows the CPU utilization of the partition along with the wait information for the collection. This perspective can often give you a good idea of the timeframes you may want to consider for deeper analysis.
Figure 7. CPU utilization and waits overview
Every chart has the following tools, shown in Figure 8, that you can use to interact with the chart:
Figure 8. Tools
There are navigation shortcuts at the top of every perspective. The Perspective and Edit navigation shortcuts provide easy ways to access the GUI controls rather than having to scroll to the bottom of the chart to use the buttons for Done, Options, and Save As.
In addition, the View option allows you to toggle on or off for additional information about the collection you are viewing. View -> Show Context shows basic information about the collection, while View -> Show System Information shows information about the system from which the collection originated. The History option displays the drill-down perspectives that have been used to get to the current perspective. In addition, the History option allows you to return to the Home page, which is a fast way to exit the main Investigate Data panel.
Figure 9. Navigation shortcuts
In any perspective, you have drill-down options to go to other charts or tables for analysis purposes. In addition to the drill-down options, there are several additional functions found in the Select Action drop-down list.
Navigator for i has a function called Set Target System. This allows the web infrastructure to support Navigator to run on one partition, but you can manage a second partition. This is a good option if you have a partition in which you cannot run a web server but wish to manage it with Navigator or use PDI to review performance data. The target system is accessed using the host servers.
If you are managing the performance of a partition, there may be times when you want to review the same set of charts on a regular basis. You can do this manually, but the Performance Data Reports make it very easy to create reports that contain the required charts. The Performance Data Reports are found under All Tasks. You must first create a report definition that describes which charts you want to include in your report; IBM supplies some sample report definitions that you can start with. After setting up your report definition, you create your performance report using that definition. Your report can be generated as a PDF, a collection of images (JPEG and PNG), or as raw data (CSV or TXT). The PDF report is very useful for sharing the information with others.
Figure 10. Performance Data Reports
In the 7.3 release, IBM added support for historical performance data and the ability to review that data with the Graph History feature. The article Viewing historical data with Graph History reviews this feature.
The Batch Model sizing tool, found under All Tasks, allows you to use Collection Services data to model the impact of hardware changes or workload growth. The article Creating a batch model reviews this feature.
The Performance Data Investigator is a powerful tool for analyzing performance data. This article has just touched upon the major features of this graphical tool. As you use PDI, you can discover that it has a great deal of function and a lot of flexibility enabling you to view your performance data and proactively manage your system’s performance.
January 28, 2019
Get the Code »
Apache SparkArtificial intelligence+
May 14, 2019
IBM Cloud PrivateIBM LinuxONE+
Back to top