In the second article in this series, I mentioned that one of the biggest improvements in performance monitoring was the addition of Data Collector Sets. Data Collector Sets are collections of performance monitor counters and system traces, that are all related to a specific purpose. In this article, I will show you how they work.
Earlier in this series, I mentioned that a lot of administrators seem to shy away from using the Performance Monitor because of its complexity. This complexity has gotten worse as time has gone on. The Windows NT version of the Performance Monitor was really similar to the Windows Server 2003 version. Both of those versions were complex in that you had to know how to interpret the data provided by the various counters. At the same time though, Windows NT had relatively few Performance Monitor counters that you had to learn about. There have been hundreds of counters created since that time though. In fact, one of the biggest challenges in using the Windows Server 2003 and Windows Server 2008 versions of the Performance Monitor is just figuring out which counters you should be using.
There are a hand full of Performance Monitor counters that are almost always of interest in just about any situation. Typically though, just looking at the more generic counters is not going to help you to gain any sort of in depth understanding of what is happening with your system. For that, you need to monitor the counters that are specifically related to the area of the system that you are trying to diagnose. But how do you choose which counters to look at? There can be hundreds of counters that are related to an individual aspect of system performance. Some of these counters are important and others are so obscure that they are pretty much only used by Microsoft’s support staff.
This is where data collector sets come into play. Data collector sets are groups of Performance Monitor counters that are specifically related to some troubleshooting task. Windows Vista and Windows Server 2008 contain four built in data collector sets; LAN Diagnostics, System Diagnostics, System Performance, and Wireless Diagnostics. You also have the ability to create your own Data Collector Sets.
Now that you know what a data collector set is, I want to show you how data collector sets work. You can access the built in data collector sets by opening the Reliability and Performance Monitor and navigating through the console tree to Reliability and Performance | Data Collector Sets | System. When you expand the System container, you will see the four built in data collector sets that I mentioned earlier. If you select one of those data collector sets, you will see the individual elements that make up the data collector set, as shown in Figure A.
Figure A When you select an individual data collector set, Windows will display the elements that make up the data collector set.
If you look at the figure above, one of the first things that you will probably notice is that out of all of the elements that make up the data collector set, only one of them is classified as a performance counter. The rest are classified as Trace or as Configuration. I will address the Trace and Configuration elements in Part 4 of this series. For now, I want to focus on the Performance Counters.
If you double click on the Performance Counter listing, you will be taken to the Performance Counter Properties sheet, shown in Figure B. As you can see in the figure, the Performance Counter Properties sheet contains multiple performance monitor counters that are to be included in the data collector set. In fact, you will notice that each of the counters that is listed is actually a wild card. For example, the first counter on the list is ‘Network Interface (*)’* This means that the data collector set includes all of the counters for all of the counters in the Network Interfaces performance object.
Figure B These are the Performance Monitor counters that are included in the data collector set.
As you can see, data collector sets help to simplify things by telling Windows which Performance Monitor counters to use, rather than requiring you to manually pick and choose. Of course you still have to be able to interpret the data that you collect, and the data collector sets can help with that too. In the next article in the series, I will show you how.
Got a question? Post it on our Windows Server 2008 forums!