Performance Monitor

Performance Monitor allows you to view your computer’s performance for such things as the Processor and Hard Disk usage. Using Performance Monitor can help identify potential bottlenecks which may be slowing a system down. It can also be used to monitor the performance of other machines on the network.

To access the Windows Server 2003 Performance Monitor click on Start.

Highlight All Programs.  Highlight Administrative Tools. Click Performance.

Another – and quicker – way of accessing the Performance monitor is by using the Run command. Click on Start. Click Run.

Type in Perfmon in the Run dialog box. Then click OK.

The Windows 2003 Performance Monitor will appear. The current window shows a chart. This chart can show multiple Counters i.e. the objects to be monitored.

Click on the Add icon to add a Counter.

Default Counters

Pages/sec: This counter measures the amount of times the page file on a hard drive is accessed. A high value of pages/sec indicates low available RAM (a value over 20 is considered high with a value of 4/5 being ideal). The solution would be to add more memory

Average disk queue length: This measures the number of operations waiting to be written to the hard drive. A high value indicates a slow disk drive (a consistent value of 2 or above is considered high). The solution would be to add a faster hard drive or split operations between multiple hard drives.

%processor time is a measure of how much work the processor is doing. A consistently high value (80% and above) indicates an overworked processor. The solution would be to either add additional processors and/or upgrade to a faster processor.

This is the Add Counters Window.

The Performance object is the actual device you want to monitor, e.g. The Processor.

The counters list allows you to monitor various settings for the Performance object you have selected. e.g. % Processor Time.

Click on the Add button to add a counter to monitor the Processor Time.

Click on the down arrow in the Performance Object to select additional Performance Objects……and a list of additional Performance objects is displayed. Click on the Memory Object.

Select the Available MBytes counter to monitor how much Memory is available. If this value drops below 4MB there may be a memory bottleneck.  Click Add and then select Close to begin Monitoring. Now the chart shows the current Processor time and Available Mbytes.  (There are also two other views, the Histogram View and the Report View.)

Click here to select the Histogram view. Click here to select the Report View. The Report view gives the most accurate reading.

The information gleaned from the performance monitor suggests that both the Processor and Memory are in working order.  It is best to monitor a machine during peak usage time.

The Processor Object

The Processor counter gives detailed reports on how the Processor is performing. Processor performance should  be monitored regularly.

The Memory Object

The Memory object gives detailed reports on how well the Memory is performing. Memory performance should be regularly monitored to determine if and when an upgrade is needed.

Monitoring Disk Performance.

Physical disk Counters are enabled in Windows 2003 by default, so this option should be available through Perfmon. Perfmon opens the Performance console which is used for setting and viewing counters.

From Start, choose Run and type in perfmon. (There are other ways of opening this utility.) Scroll to find the Physical Disks option. Selecting the object reveals a choice of counters. For the sake of illustration, two disk queue counters will be selected.

First the Average disk write queue length is selected. Secondly the Current Disk Queue Length is selected. The two traces are displayed now in the right pane.  This value suggests that the present disks configuration can cope with the current usage. Remember the caveat about sampling during times of peak usage e.g. when the most read/write operations are occurring.

The option to add counters to monitor Logical Disk performance is available. Once Logical Disk is selected, a particular counter must be chosen: This is a useful tool, for a brief description of the function of the selected counter.

If there is more than one logical drive  available then you will be able to select it here. For illustration, the C partition will be chosen.

This display shows  that the drive is in first class shape. If the visible peak is a sustained feature however, then this is a sign of fragmentation.

Additional Monitors

NTDS Is used for monitoring the Active Directory. By using System Monitor you can track the performance of the active directory database.

Domain Name System (DNS) Monitoring is used to check and troubleshoot DNS servers, also to troubleshoot the DNS configuration for Active Directory.

Dynamic Host Configuration Protocol (DHCP) Monitoring is used for troubleshooting and monitoring the performance of DHCP Servers.

Identifying and Troubleshooting Bottlenecks

Bottlenecks can be caused by resources not being used efficiently, a resource being too slow or too small, or the overuse of memory intensive applications.

By using Performance Monitor, the performance of various parts of the system can be measured allowing system bottlenecks to be identified.

It is not enough to view one set of counters and pronounce that a disk or processor cannot cope in its present configuration. For example, processor bottlenecks can be caused by memory bottlenecks, “unintelligent” network cards, or even outrageously demanding screen savers!

Often one set of counters must be compared with another to get a truer picture. For example, the % Disk Time counter gives a reading which needs to be compared with the % Idle Time counter to give a meaningful result.

Various counters can be considered together to detect possible conflicts. Even though the problem may appear to be with one device it may lie with another entirely.

If Processor is above 80%, then there may be a problem with the processor, or the processor is too slow for the applications that are running on the system.

If there is a high split I/O sec rate on the drive then the disk might need defragmenting.

If there is a high %page file usage then the addition of more memory is indicated.

Lack of memory is the most common cause of performance problems in your Windows computer. Physical memory installed can be identified by accessing the System Properties page after  right clicking My Computer then selecting Properties.

Paging File

The Paging File  – also known as the Swap File/Virtual Memory – is a file on the hard drive that acts as  temporary memory space when the physical memory is full. A good way to troubleshoot paging file problems is to move the paging file to a separate physical disk and IDE channel (Or a separate hard drive for SCSI systems). This is good practice in any event.

Windows uses dynamic page file sizing. The continual re-sizing of the page file, as demand increases and decreases, puts additional load on system resources. By setting the minimum size to the maximum size, the page file stays at a constant size regardless of the system usage. This lowers the overhead on the processor and the hard drive.

If performance with the Swap File becomes an ongoing problem, a better long-term solution may be to increase a computer’s memory.

To amend the size of the paging file, right click My Computer and select Properties. Select the Advanced tab. Click on Performance Settings.

The visual effects tab allows you to tune the machine to enhance performance by reducing visual effects. Both processor and memory can be tuned to enhance server function or application function. Click on Change to open the Virtual Memory window. To change the location of the page file select a different drive and set the size required. Then on the original drive select no paging file. After every single change click the Set button otherwise your settings will be forgotten by the machine.