This chapter explains how to use the ServerStats features of Oracle Load Testing to monitor systems and metrics during a load test of the application-under-test.
ServerStats lets you monitor a variety of server-side application, database, system, and Web server statistics. You can configure ServerStats to display real-time performance statistics for the various hosts and services available from the server, such as percentage of CPU usage, memory usage, Web server statistics, etc.
The ServerStats configurations contain one or more configured monitors for monitoring performance of application infrastructure during a load test. Once saved, you can add monitors into the configuration and apply them to the system to be monitored.
The Metric Profiles and Metrics options in the ServerStats options are used to create generic metric profiles and metrics that can be saved for use in multiple configurations. These generic definitions are available when you add a monitor to the configuration. Metric definitions monitor individual data points. Metric profiles monitor a group of data points.
Metric Profiles and Metrics are shipped with a default set of metrics. Additional metrics not included in the default metrics can be accessed by creating custom profiles or metrics.
Oracle Load Testing supports the following metric profiles:
Systems
WebServers
Refer to Appendix E for a description of the metrics included with each profile.
Remote data collectors can be configured to run ServerStats monitors from remote locations.
To create a ServerStats monitor, the following broad tasks must be completed:
Configure the target systems you want to monitor and the data collectors you are using either in the Options or directly from the ServerStats Monitoring configuration.
Create generic metric profiles and metrics if you want to use them in more than one configuration, otherwise metric profiles and metrics can be added directly from the configuration.
Create and save a configuration.
Add monitors to the configuration.
Refer to Appendix E for examples.
ServerStats lets you maintain multiple configurations of servers, systems, and counters in different configuration files that can be selected in the Monitoring section of the Composer tab to be run with a load test. Configuration files store a particular set of ServerStats settings for monitoring server performance under load. The files store the following:
Configured data sources
Selected counters from each data source
Selected metric profiles and metrics
Data source and counter properties
Configurations are listed in the Monitoring section of the Composer tab.
Note:
ServerStats configurations are only initiated and run as part of a load test. If you need to add additional metrics during a load test, start a new session and compare data later using absolute time on the reports.To create a new configuration file:
Select Monitoring on the Composer tab.
Click New to clear any existing configuration.
Click Add Target to list the system types.
Select the system type:
Database - opens a dialog box for specifying a database system to monitor.
Weblogic - opens a dialog box for specifying a Weblogic system to monitor.
Linux - opens a dialog box for specifying a Linux system to monitor.
Windows - opens a dialog box for specifying a Windows system to monitor.
Generic Target - opens a dialog box for specifying any system to monitor.
Existing - opens a list of existing target configurations that have already been created.
Specify the system connection information for the target system type.
If you want to use the default profile counters for the system type, make sure the Apply [system type] profile automatically option is selected.
Click OK.
If you do not use the Apply [system type] profile automatically option or you want to add additional counters to the configuration, click Add Counters and select where you want to retrieve the counter from and add the counter.
Click Test to make sure the monitors are configured correctly. Oracle Load Testing tries to start the monitors and collect data from the systems the monitors are configured against.
Click Save Configuration.
Enter a name for the configuration file and click OK.
To rename a configuration:
Select Options from the User menu.
Select Configurations in the ServerStats group to open the Configurations options.
Select the configuration you want to rename.
Click Edit above the list of configurations to display the Edit Configuration dialog box for changing the name of the configuration.
Name - displays the name of the configuration.
Edit the name.
Click OK to exit the Configurations dialog box.
To edit a saved configuration file:
Select Monitoring on the Composer tab.
Click Load configuration.
Select the configuration file to load:
Use the Add Counter menu to add counters to the configuration.
Use the Edit option to change the properties of an existing monitor in the configuration.
Use the Delete option to remove a counter from the configuration.
Click Test to make sure the monitors are configured correctly. Oracle Load Testing tries to start the monitors and collect data from the systems the monitors are configured against.
Click Save Configuration.
You can add the following data source monitors:
COM+ - The ComPlus Monitor retrieves data for counters provided by ComPlus objects on Windows 2000 or later operating systems. The counters that are available are determined by the ComPlus applications that are installed on your system.
Database - You can set up a monitor that performs a database query for an individual machine.
JMX - The JMX monitor lets you determine if a Java application is performing properly. The JMX monitor checks the specified JMX MBean Attribute values to confirm that the Java application is performing as expected. The JMX monitor also provides an easy way to confirm whether or not the JMX application is running.
Perfmon - A Perfmon monitor checks on the performance of a specific Perfmon metric on a single machine. The monitor must have appropriate privileges to access the Perfmon statistics on the monitored machine.
Ping - Ping is an easy way to confirm that a host is responding The Ping monitor sends an echo request to a specified host and waits for a reply. The monitor returns the round trip time.
URL - A URL monitor lets you determine if a URL that is critical to your web site is functioning as it should. In addition, the URL monitor can be used to parse data from the Apache server status page using a regular expression.
Virtual Agent - You can set up Virtual Agent monitors to run an executable program from the command line of the system. You specify a Perl expression to return metrics from the command. Virtual Agent is an easy way to execute operating system or custom programs on a host and return specific metrics. The Virtual Agent used to be called Remote Agent.
IBM Websphere PMI - IBM Websphere PMI is an easy way to confirm that a host is responding.
To monitor disk performance on a Windows machine, Diskperf -yv must be enabled on that machine. See "Using Diskperf" later in this chapter for more information.
Select the Options section on the User menu.
Select Metrics in the ServerStats group.
Select New and select the data source for the type of monitor you want to add.
Enter the data required to configure the monitor. All monitors include the following properties:
Name - enter a name for the monitor.
Description - enter a description for the monitor.
Category - displays the metric's category.
Data Source - displays the metric's Data Source type.
Translation Type - Translation Type controls how a metric's data is displayed in Reports.
No Translation - data is displayed as it is provided by the metric's data source.
Rate (minutes) - data is presented as a number, per minute. For example, if the time interval between two samples was 10 minutes, and 50 errors were found, the Rate (minutes) would be five (5) errors per minute.
Note:
Rate (minutes) translates data sources that return their values as an endlessly increasing counter. If the data source does not return data this way, use No Translation.SampleFrequency - specifies how often to sample data.
Sample Multiplier - enter the multiplier you want to be applied to the data that is collected. Setting the sample multiplier to 1 means that the data will be reported as is. If the metric measurement is expressed in kilobytes, setting the sample multiplier to 1024 translates it to bytes. If the metric measurement is expressed in milliseconds, setting the sample multiplier to .001 translates it to seconds.
Click OK.
The following sections list the additional properties required for specific data source types.
This section describes the additional fields required for the COM+ data source monitor.
Object ID - displays the object ID.
Object Name - displays the object name.
This section describes the additional fields required for the Database data source monitor.
Query - Enter a valid database statement. For example, you can enter the following Oracle SELECT statement:
select value from v$sysstat where statistic# = 1
If your database server supports returning a result set from a stored procedure you can call the procedure. For example, the following SQL Server stored procedure, present in the database, returns a result set (not an output parameter) showing the configured sample rate (samples/minute):
p_computeSampleRate
Report Type
Measure Query Time - Select this option to see how long the query takes.
Read Result Set - Select this option to match a value in the result.
Result Pattern - Enter the Perl expression for the value that should be returned for the specified query.
This section describes the additional fields required for the JMX data source monitor.
MBean Name - the fully qualified MBean name.
Attribute Name - the MBean attribute name to request.
Sub Attribute Name - the MBean sub attribute name to request.
Description - enter a description for the MBean.
Filter MBeans By - select a filter to use.
Discovery MBeans - lists the available MBeans.
This section describes the additional fields required for the Perfmon (Windows Performance Monitor) data source monitor.
Object - specifies the performance object name.
Counter - specifies the counter name.
Instance - specifies the instance of the counter to monitor.
This section describes the additional fields required for the Ping data source monitor.
Measurement - Specifies the metric to use for the Ping monitor. Select the metric to use: Average Round Trip Time (PingAvgRTT) or Percentage of Success (PingPercentSuccess).
Length - Specifies how many bytes should be included in each echo request packet.
Time to Live - Specifies the maximum number of router hops permitted. If the echo request takes more than the maximum number of hops, it is thrown away and the Ping attempt is treated as a failure.
This section describes the additional fields required for the URL data source monitor.
URL Usage - specifies whether to use the URL specified in the System Homepage field on the Add/Edit Monitored System dialog box.
Use System Homepage - use the URL specified in the System Homepage field on the Add/Edit Monitored System dialog box.
Specify - enables the URL field for specifying the URL of the page you want to monitor.
URL - Type the URL for the web page you want to check. The URL must be specified with the full protocol designation of http or https. For example: http://www.company.com.
If you use Get as the HTTP Method (the default), simply include any additional parameters you want to include as part of the URL string, for example:
http://search.yahoo.com/bin/search?p=electronics
POST - the Post method, usually associated with a form, allows a much larger amount of information to be associated with the request.
GET - include any additional query strings as part of the URL address.
Note:
Some servers limit the length of URL and query strings to about 240 characters.Post Data - If you select Post as the URL Method, you can add a series of name-value pairs. For example, a URL generated after completing a form may appear as follows:
POST http://www.company.com/cgi-bin/guestbook.pl ...(additional headers) name=Harold&nickname=Harry
To reference this page, enter name=Harold&nickname=Harry in this field.
User Agent - you can specify a particular user agent when you monitor a URL. When you configure a User Agent, the browser and operating system you specify is added to the header of the URL request. This allows you to test pages designed for a specific browser URL confirm that a page designed for multiple browsers works appropriately for each of them.
Enable Cookies - specify if cookies are enabled.
False - select this option if you want the URL test to prevent the use of cookies when accessing the page.
True - select this option if you want to allow the URL monitor to support cookies as part of the URL request. A fresh cookie is retrieved each time the test is run.
Custom Cookie - type in a cookie value for the page you are monitoring. If you supply a particular value, remember that the cookie is saved and can become out of date. Refresh cookie requests are ignored.
Include image download times - calculates the URL availability and access time with retrieval of embedded images (such as .gif and .jpeg files and references within <img> tags).
Include object download times - calculates URL availability and access time with retrieval of embedded objects (such as ActiveX objects and Java applets). Oracle Load Testing measures the time taken to retrieve data used by objects, but does not include the time needed to retrieve any executable file required in order to interpret the data. Content types other than HTML and images are also classified as objects.
Result of Perl Expression - enter the Perl expression to use if you have selected Result of Perl Expression as the metric type. For example, you could use the following Perl expression to extract requests per second from the Apache Server status page:
/([0-9]+\.[0-9]) requests\/sec/
Username - if the URL specified requires a name and password for access, enter the name in this field.
Password - if the URL specified requires a name and password for access, enter the password in this field.
Use HTTP Proxy - select True if your network makes use of a proxy server.
Proxy Host - specify the name or IP address of the proxy server.
Proxy Port - the port of the proxy server.
Proxy Username - the proxy server login user name.
Proxy Password - the proxy server login password.
This section describes the additional fields required for the Virtual Agent data source monitor.
Command Line - specifies the command to execute on the host machine.
Matching Regexp - specifies a Perl expression to use to report the data returned from the executable program.
Key of value in Regexp - specify the key of value to return from the Regular Expression specified in Matching Regexp. Key of value 0 returns the first match, 1 returns the 2nd match, etc. The match is of the Regular Expression inside the parenthesis (). For example, if the Command Line is free -m and the Matching Regexp is /Mem:.+?(\d+).+?(\d+)/, a Key of Value 0 returns the value in the first (\d+) of the Regular Expression.
This section describes the additional fields required for the IBM Websphere PMI data source monitor.
Document - specify the WebSphere document.
Server - specify the server instance name that you want to connect with.
Metric profiles are collections of metric definitions for monitoring general system types such as WebSphere, Oracle, etc. A profile can then be used to monitor a specific server when it is added to a ServerStats configuration.
Select Options from the User menu.
Select Metric Profiles in the ServerStats group.
Click New.
This dialog box is where you select the metric to include in the metic profile you want to add. Expand the appropriate node and select the data source, metric, or metric profile that you want to add.
Expand the Available Counters node, select the metric you want to add and click Add.
Repeat step 4 for each metric you want to add to the Metric Profile.
Click OK when finished
To create monitors using metrics:
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New.
Select the Data Source Type..
Specify the Metric properties.
Click OK when finished.
To edit a monitor in a configuration:
Select Monitoring on the Composer tab.
Select Load Configuration if you have a save configuration file.
Select the configuration you want to edit and click OK.
Select the monitor that you want to edit and click Edit. .
Make any changes.
Click OK when you are done.
This section explains how to work with metric profiles. Metric Profiles are a predefined group of metrics that can be added to a monitoring configuration by adding the Metric Profile rather than the individual metrics.
You can add a user defined metric profile that contains any combination of default and user defined metrics.
Select Options from the User menu.
Select Metric Profiles in the ServerStats group.
New - displays the Add Profile dialog box for adding a user defined metric profile.
Edit - displays Edit Profile dialog box. This option is only available for user defined metric profiles. Default profiles cannot be edited.
Clone - displays the Clone Profile dialog box for copying a default profile to create a user defined metric profile that can be edited.
Delete - deletes the selected user defined metric profile. Default profiles cannot be deleted.
Import - opens a dialog box for selecting the Metric Profile .zip file to import.
Export - opens a dialog box for selecting the Metric Profile .zip file to import.
Metric Profile List - lists the default profiles and user defined metric profiles.
Click New. The Add Metric Profile dialog box is displayed.
Name - the name of the profile.
Description - the profile's description.
Category - displays the category to which the profile belongs.
Available Counters - displays the list of available metrics.
Profile Metrics - lists the selected metrics.
Add - adds the selected metric to the Profile Metrics list.
Remove - removes the selected metric from the Profile Metrics list.
Enter a name and description for the profile.
Add metrics by selecting them in the Available Counters list and clicking Add.
Click OK when finished. The new profile is listed in the User Defined category in the list of metric profiles.
You can edit user defined metric profiles. Default profiles cannot be changed. Changing a metric profile does not change it in any configurations using the metric profile.
Select Options from the User menu.
Select Metric Profiles in the ServerStats group.
Select the user defined metric profile that you want to edit and click Edit. The Edit Profile dialog box is displayed.
Name - the name of the profile.
Description - the profile's description.
Category - displays the category to which the profile belongs.
Available Counters - displays the list of available metrics.
Profile Metrics - lists the selected metrics.
Add - adds the selected metric to the Profile Metrics list.
Remove - removes the selected metric from the Profile Metrics list.
Add metrics by selecting them in the Available Counters list and clicking Add.
Remove metrics by selecting them in the Profile Metrics list and clicking Remove.
Click OK when finished.
Cloning a profile lets you start with a default profile and create a user defined metric profile that you can customize for your needs by adding or deleting metrics.
Select Options from the User menu.
Select Metric Profiles in the ServerStats group.
Select the profile that you want to clone and click Clone. The Clone Metric Profile dialog box is displayed.
Name - the name of the profile.
Description - the profile's description.
Category - displays the category to which the profile belongs.
Available Counters - displays the list of available metrics.
Profile Metrics - lists the selected metrics. Click the "x" in front of a metric to delete it from the list.
Add - adds the selected metric to the Profile Metrics list.
Remove - removes the selected metric from the Profile Metrics list.
Enter a name for the profile.
All of the metrics for the selected profile are automatically added to the Profile Metrics list.
Add additional metrics by selecting them in the Available Counters list and clicking Add.
Remove metrics by selecting them in the Profile Metrics list and clicking Remove.
Click OK when you are done. The new profile is listed in the User Defined category in the list of metric profiles.
Only user-defined metric profiles can be deleted. To delete a user-defined metric profile:
Select Options from the User menu.
Select Metric Profiles in the ServerStats group.
Select the user defined metric profile you want to delete and click Delete.
Click Yes when asked to confirm the deletion.
To test metric profiles:
Select Monitoring on the Composer tab.
If you have a saved configuration with a Metric Profile select Load Configuration and open the configuration file.
If you do not have a saved configuration, click Add Target and select the system type or existing target.
Click Add Counters.
Select From Library.
Select the Metric Profiles tab.
Select the Metric Profile and click Add.
Click OK.
Click Test.
This section explains how to work with metrics.
You can add metrics that can be used in any configuration. Metrics that you add are listed in the User Defined node of the Metrics tree. The Metrics dialog box has the following options:
Select the Options section on the User menu
Select Metrics in the ServerStats group.
Select New and select the data source for the type of metric you want to add.
Enter the data required to configure the metric. All monitor include the following properties:
Name - enter a name for the monitor.
Description - enter a description for the monitor.
Category - displays the metric's category.
Data Source - displays the metric's Data Source type.
Translation Type - Translation Type controls how a metric's data is displayed in Reports.
No Translation - data is displayed as it is provided by the metric's data source.
Rate (minutes) - data is presented as a number, per minute. For example, if the time interval between two samples was 10 minutes, and 50 errors were found, the Rate (minutes) would be five (5) errors per minute.
Note:
Rate (minutes) translates data sources that return their values as an endlessly increasing counter. If the data source does not return data this way, use No Translation.SampleFrequency - specifies how often to sample data.
Sample Multiplier - enter the multiplier you want to be applied to the data that is collected. Setting the sample multiplier to 1 means that the data will be reported as is. If the metric measurement is expressed in kilobytes, setting the sample multiplier to 1024 translates it to bytes. If the metric measurement is expressed in milliseconds, setting the sample multiplier to .001 translates it to seconds.
The following sections explain how to add specific metric types:
Select Options from the User menu.
Select the Metrics in the ServerStats group.
Click New.
Select ComPlus and specify the metric properties. ComPlus metrics require the following properties.
Object Name - enter the object name of the counter.
Object ID - enter the object identifier.
Specify the metric properties and click OK.
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New.
Select Database and specify the metric properties. Databse metrics require the following properties.
Query - Enter a valid database statement. For example, you can enter the following Oracle SELECT statement:
select value from v$sysstat where statistic# = 1
If your database server supports returning a result set from a stored procedure you can call the procedure. For example, the following SQL Server stored procedure, present in the database, returns a result set (not an output parameter) showing the configured sample rate (samples/minute):
p_computeSampleRate
Measure Query Time? - Select true to see how long the query takes. Select false to match a value in the result.
Read Result Set? - Select true to return a result set if your database supports it. Select false to return an output parameter.
Result Pattern - enter the Perl expression for the value that should be returned for the specified query.
Enter the Metric properties.
Click OK.
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New.
Select JMX and specify the metric properties. JMX metrics require the following properties:
MBean Name - specify the fully qualified MBean name.
Attribute Name - specify the MBean attribute name.
Sub Attribute Name - specify the MBean sub attribute name.
Description - enter a description of the metric.
Filter MBeans By - select the server type by which you want to filter MBeans or select All Counters to return all supported MBeans.
Discovery MBeans - select the type of MBeans to discover or select All Counters to discover all types of MBeans.
Enter the metric properties and click OK.
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New and select Perfmon.
Select Perfmon and specify the metric properties. Perfmon metrics require the following properties.
Object Name - specify the object name. For example, Processor or Memory. You must enter the object name exactly as it is displayed in the Perfmon editor. Be careful to check for spaces and case in the Object name.
Counter Name - specify the counter name. For example, % Processor Time. You must enter the counter name exactly as it is displayed in the Perfmon editor. Be careful to check for spaces and case in the Counter name.
Instance Name - if the metric has an instance associated with it, specify it here.
Enter the Metric properties.
Click OK.
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New.
Select Ping and specify the metric properties. Ping metrics require the following properties.
Measurement - Specifies the metric to use for the Ping monitor. Select the metric to use: Average Round Trip Time (PingAvgRTT) or Percentage of Success (PingPercentSuccess).
Length - specify how many bytes should be included in each echo request packet.
Length - specify how many bytes should be included in each echo request packet.
Enter the Metric properties.
Click OK.
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New.
Select URL and specify the metric properties. URL metrics require the following properties.
Specify - select this option if you want to specify a URL. Enter the URL in the URL field.
Use System Homepage - (default) select this option to automatically use the home page defined for the system.
URL - If URL Usage is set to Specify, type the URL for the web page you want to check. The URL must be specified with the full protocol designation of http or https. For example: http://www.company.com.
If you use Get as the HTTP Method (the default), simply include any additional parameters you want to include as part of the URL string, for example:
URL - If URL Usage is set to Specify, type the URL for the web page you want to check. The URL must be specified with the full protocol designation of http or https. For example: http://www.company.com.
If you use Get as the HTTP Method (the default), simply include any additional parameters you want to include as part of the URL string, for example:
http://search.yahoo.com/bin/search?p=electronics
POST - the Post method, usually associated with a form, allows a much larger amount of information to be associated with the request.
GET - include any additional query strings as part of the URL address.
Note:
Some servers limit the length of URL and query strings to about 240 characters.Post Data - If you select Post as the URL Method, you can add a series of name-value pairs. For example, a URL generated after completing a form may appear as follows:
POST http://www.company.com/cgi-bin/guestbook.pl ...(additional headers) name=Harold&nickname=Harry
To reference this page, enter name=Harold&nickname=Harry in this field.
User Agent - you can specify a particular user agent when you monitor a URL. When you configure a User Agent, the browser and operating system you specify is added to the header of the URL request. This allows you to test pages designed for a specific browser URL and confirm that a page designed for multiple browsers works appropriately for each of them.
False - select this option if you want the URL test to prevent the use of cookies when accessing the page.
True - select this option if you want to allow the URL monitor to support cookies as part of the URL request. A fresh cookie is retrieved each time the test is run.
Custom Cookie - type in a cookie value for the page you are monitoring. If you supply a particular value, remember that the cookie is saved and can become out of date. Refresh cookie requests are ignored.
Types other than HTML and images are also classified as objects.
Include image download times - calculates the URL availability and access time with retrieval of embedded images (such as .gif and .jpeg files and references within <img> tags).
Include object download times - calculates URL availability and access time with retrieval of embedded objects (such as ActiveX objects and Java applets). Oracle Load Testing measures the time taken to retrieve data used by objects, but does not include the time needed to retrieve any executable file required in order to interpret the data.
Include object download times - calculates URL availability and access time with retrieval of embedded objects (such as ActiveX objects and Java applets). Oracle Load Testing measures the time taken to retrieve data used by objects, but does not include the time needed to retrieve any executable file required in order to interpret the data.
Perl Expression - enter the Perl expression to use if you have selected Result of Perl Expression as the metric type.
Username - if the URL specified requires a name and password for access, enter the name in this field.
Password - if the URL specified requires a name and password for access, enter the password in this field.
Enable Proxy- select this checkbox if your network makes use of a proxy server.
Proxy Host - specify the name or IP address of the proxy server.
Proxy Port - the port of the proxy server.
Proxy Password - the proxy server login password.
Enter the Metric properties.
Click OK.
To add a Virtual Agent metric:
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New.
Select Virtual Agent and specify the metric properties. Virtual Agent metrics require the following properties.
Command Line - specifies the command to execute on the host machine.
Matching Regexp - specifies a Perl expression to use to report the data returned from the executable program.
Key of value in Regexp - specify the key of value to return from the Regular Expression specified in Matching Regexp. Key of value 0 returns the first match, 1 returns the 2nd match, etc. The match is of the Regular Expression inside the parenthesis (). For example, if the Command Line is free -m and the Matching Regexp is /Mem:.+?(\d+).+?(\d+)/, a Key of Value 0 returns the value in the first (\d+) of the Regular Expression.
Enter the Metric properties.
Click OK.
To add an IBM Websphere PMI metric:
Select Options from the User menu.
Select Metrics in the ServerStats group.
Click New.
Select IBM Websphere PMI and specify the metric properties. WebSphere metrics require the following properties.
Document - specify the URL to use as the root for the queries to identify counters. For example, wasPerfTool/servlet/perfservlet.
Server- specify the server instance that you want to talk to.
Enter the Metric properties.
Click OK.
Only user-defined metrics can be edited. Default metrics cannot be changed. Changing a metric changes it in any configuration that uses it.
To edit user-defined metrics:
Select Options from the User menu.
Select Metrics in the ServerStats group.
Select the user defined metric you want to edit and click Edit. The Edit Metric dialog is displayed showing the properties pertinent to the selected metric.
Make any necessary changes.
You can test the metric by clicking Test and specifying the test parameters.
Click OK.
When you clone a default metric you create a user defined metric that can be edited, saved, and used in configurations.
To clone a metric:
Select Options from the User menu.
Select Metrics in the ServerStats group.
Select the metric you want to clone and click Clone. The Clone Metric dialog box is displayed. This dialog box displays name and description details about the metric as well as metric discovery and property information when applicable.
Make any changes and click OK. The metric is listed in the user defined node of the metric tree.
User defined metrics can be deleted. Default metrics cannot be deleted.
To delete a metric:
Select Options from the User menu.
Select Metrics in the ServerStats group.
Select the user defined metric that you want to delete and click Delete.
Click Yes when asked to confirm the deletion.
To test a metric:
Select Monitoring on the Composer tab.
If you have a saved database configuration file select Load Configuration and open the configuration file.
If you do not have a saved configuration, click Add Target and select the system type. Enter the target system properties, click OK and add counters to the configuration.
Click Test.
When the test finishes, click OK to exit the Test Monitoring Configuration window.
You can view metric descriptions and configuration details from Monitoring section of the Composer tab.
To view metrics details:
Select Monitoring on the Composer tab.
If you have a saved database configuration file select Load Configuration and open the configuration file.
If you do not have a saved configuration, click Add Target and select the system type. Enter the target system properties, click OK and add counters to the configuration.
Click Edit next to the metric. The Edit Monitor dialog box show details about the metric and the Data Collector system.
Click OK to exit the Edit Monitor dialog box.
Diskperf is a command that enables the Windows Perfmon counter that measures computer disk performance.
You will probably have to enable Diskperf on Windows machines when you have monitors that use Perfmon. These counters are may not be enabled by default on a Windows machine.
To execute the Diskperf command:
Log on with administrative rights to the NT or Windows 2000 machine whose disk performance you want to monitor.
At the command line, type one of the following:
Non-RAID or hardware RAID PCs: diskperf -y
Software RAID PCs: diskperf -y e
Windows 2000 PCs: diskperf -yv
Note:
The complete list of options for either system can be found by runningdiskperf -?Press Enter.
Restart the computer.