Performance Monitor- Perfmon

                            Collecting and Reading Performance Information in Performance Monitor

This technique can also be used to debug other performance issues and monitor performance-related items such as disk usage, memory usage, page usage and processor usage.

Start the ASA database server

Start > Run… 

Type: perfmon 

This starts the Performance Monitor:

Right click on Counter Logs and select New Log Settings … 

Enter an appropriate name (i.e. Sample System Performance),

    Click on OK. 

This opens a Sample System Performance interface, with a current log file name of 

C:\PerfLogs\Sample_System_Performance_000001.blg under the General tab

Select the Add Objects… button, and you will be prompted with an Add Objects interface. 

Leave the default Select counters from computer: \\<YOURPCNAME> radio button, then select each of the Performance objects (you should see all highlighted) – select the first item, hold down shift and click on the last item to select all. (It is unfortunate that Microsoft did not give the option to select all. Please note, the items to be chosen depend on what is installed to the system you are running Performance monitor on.

Select the Add button. 

You will now see the selected items listed in the counters box. For example:

Note, you may want to change the default interval from the default 15 seconds to something with a lighter impact on the server. To do this, change the Interval entry from 15 to 5 and change the Units pull-down from seconds to minutes. 

Under the Log Files tab, you can enter a comment indicating the purpose for this information being collected, although this is not required:

Under the Schedule tab, the schedule can be changed from a scheduled Monitor start-up (Automatic) to a Manual start/stop Monitor:

Select the Apply button – this will start the logging. Click on OK. 

In the above example, you should see the following in the Performance interface

And, if you double click on the sample just created you will see the following:

Right-click on the sample just created in Performance interface, and click on the Start Button 

If not already running, run your application(s). 

Right click on the log file when you are finished your sample period and select Stop.

To locate the Performance Monitor .blg Output File

If you need to send the output file to someone else for analysis, then go to the C:\PerfLogs\ directory, and select the respective .blg file. (in this example, the Sample_System_Performance_000001.blg ).

To View a Performance Monitor .blg Output File

In Performance Monitor, select the System Monitor, the click on the View Log File Data icon (looks like a database drum). This will prompt for a source:

Change the Data source from “Current activity” to “Log files” and select the Add button. This will open a “Select Log File” interface that allows you to browse to your log file location. When you find the appropriate log file highlight it. For example, in the Sample_System_Performance_000001.blg file created earlier, you will see the resulting .blg file in the C:\PerfLogs\ directory.

Click the Open button:

Click on the respective file, then the Apply button. Note, you will see entries are populated for the Total range:

Click the OK button. 

Select the view chart icon:

You can customize this information to drill down to helpful information. You can remove superfluous lines by highlighting them and pressing delete.

Viewing Graphs in Performance Monitor

In most cases you can ignore spikes in the counters. If you see high values in a particular counter at a certain time, see if the counter is being influenced by SQL Server or by outside factors such as backup software, virus scanners, etc. If it is definitely tied to SQL Server, you can match the time span for the counter to queries executed in Profiler during that timeframe and work on tuning them.

Plan and Implement Changes

Test your changes on a representative test system. This means similar hardware, configuration, data, statistics, etc. Then schedule a rollout on your production system. Create new logs to track the improvement made by implementing the changes. Use these logs, along with documentation regarding data size, system load, etc., as a new baseline for future performance comparisons.