Measuring Reliability and Performance
Knowing how your system is performing is the first step in locating problems of all sorts. Performance and reliability information can help you assess both the hardware (physical) and software status of your system. It can also help you locate problem areas and even predict when failures might occur based on trends. Of course, before you get any of this information, you must actually monitor the system by using the Reliability and Performance Monitor console.
The initial display you see provides you with a quick overview of the resources on your server. In many cases, this overview is all you need to determine whether you need to dig further into the system performance information. At some point, you begin to recognize how your system normally performs and you can use this display to see when something is wrong. When you do suspect a problem, you likely need to dig further using the features in the Performance Monitor and Reliability Monitor folders, described in the following sections.
Using the Performance Monitor
The Performance Monitor presents a display of performance information as you configure it. The focus of this display is the graph, which contains colored lines for each of the items you're monitoring. The Performance Monitor also provides a histogram and report view, which you can use when the line graph doesn't work. Press Ctrl+G to switch between the graph types. Below this graph are simple statistics that tell you about maximum, minimum, and average values. Finally, you see a legend that describes what the display is monitoring.
The basis of the Performance Monitor is very simple. The system installs a special piece of software called a counter. A counter does precisely what its name implies - it counts something. The counter may count the number of times an event occurs per second or calculate the percentage of memory used over time. However, the idea is to monitor the system and report on a particular event over a specific time interval. The system includes hundreds of counters, and you'll find that many applications include them as well.
To remove one of the counters from the list, highlight its entry in the list and press Delete. To highlight a counter so that you can see it better, highlight it in the list and press Ctrl+H. To add a new counter, you press Ctrl+I. In this case, you see the Add Counters dialog box.
The entire display is organized to make finding a particular counter easier. You begin by selecting a computer. Depending on your network setup, you can actually monitor the same counter from several computers so that you can create a centralized display of information.
After you select a computer, you choose a performance object on that computer. The object contains a number of counters. These counters describe the performance characteristics of the object. You can select all the counters for an object by clicking the object, or you can click the plus sign next to the object to display individual counters and choose those individual counters from the list.
A counter has to measure something. In some cases, the counter has only one item it can measure, but this is the exception rather than the rule. In this case, the counter can measure the performance for all hard drives or a particular hard drive. A single measure, such as _Total or C:, is an instance of the counter.
When you finish choosing the computer, object, counter, and instance, click Add. The Performance Monitor adds the new counter to the list of counters in the left pane. You can add as many counters as needed to monitor your system for a particular need. Of course, too many counters make the display hard to read, so you have to balance your need for information against the readability of the display. When you finish adding counters, click OK and you see the new counters displayed on-screen.
Using the Reliability Monitor
The Reliability Monitor is more of a static display than the Performance Monitor. It's a kind of specialized log. The display begins with a stability report. The stability of your server is rated from 1 to 10, with higher numbers providing better stability. The stability for the selected day appears to the right of the chart. You can choose a particular date by clicking its entry.
The chart also includes little boxes that can contain a number. For example, if you uninstall a piece of software on a particular date, you see the number 1 in the box where the Software (Un)Installs entry crosses the date of interest. You can use this tabular data to determine which of the dates to select. When you find a date with an entry, you can see the details of the event in the report that appears below the chart.
The report includes a number of entries that help you quickly determine how the event affects your system. For example, a software installation includes the name of the software, version, activity (install or uninstall), activity status, and date of the event. You can usually find additional information about the event in the event logs.
In this tutorial:
- Windows Server 2008 Standard Maintenance
- Interacting with the System Applet
- Configuring Your User Interface for Maximum Functionality
- Defining the Internet Options settings
- Defining the personalization settings
- Defining the Problem Reports and Solutions settings
- Defining the Taskbar and Start menu settings
- Understanding How UAC Affects Maintenance Tasks
- Measuring Reliability and Performance
- Protecting System Data
- Performing a system restore
- Performing Disk Management Tasks
- Performing disk management
- Automating Diagnostic Tasks with Task Scheduler
- Working with Remote Desktop
- Accessing local resources
- Creating a Windows Recovery Disc