Skip to content
CA Unified Infrastructure Management Probes
Documentation powered by DocOps

ntperf IM Configuration

Last update September 24, 2018

The Performance Collector (ntperf) probe is configured to create a profile to monitor a performance object and its respective counters and instances on a Windows system. The probe sends alarms on unexpected values and generates quality of service (QoS) messages.

Note: The IM configuration of the probe supports only English language.

This article is for probe version 2.0 or later.

The following diagram outlines the process to configure the probe.

Configuring ntperf (IM)

Contents

Verify Prerequisites

Verify that required hardware and software is available before you configure the probe. For more information, see ntperf (Performance Collector) Release Notes.

Configure General Properties

You can configure the logging and global monitoring properties of the probe through the Setup Tab.

Follow these steps:

  1. Click Setup Tab and update the following information:

    • Check Interval (Seconds): specifies the time interval at which the performance values are sampled.
      Default: 60

      Note: Reduce this interval to generate alarms and QoS frequently. A shorter interval can also increase the system load.

    • Default message level: specifies the default message level for alarm messages when you delete the sample profiles and alarms in the probe. If no alarms are specified in a profile, the probe uses the specified severity specified for the generated alarms.
      Default: Information 

    • Configurator request timeout: defines the time, in seconds, during which the configuration request remains active. This value defines the available time when retrieving current value of an object on the profile name > Alarm on Value tab.
    • Send short name for QoS source: allows you to send only the host name for the QoS source, when this option is selected. If not selected, the probe sends the fully qualified domain name.
      Default: Not selected 

    • Log File: defines the log file name.
      Default: ntperf.log
    • Log File Size: specifies a maximum size of the probe log file. Upon reaching this size, the contents of the file are cleared.
    • Log Level: specifies the level of details that are written to the log file. You can select the following log levels:

      • 0 - Logs only severe information (default)
      • 1 - Logs error information
      • 2 - Logs warning information
      • 3 - Logs general information
      • 4 - Logs debugging information
      • 5 - Logs tracing/low-level debugging information

      Note: Log as little as possible during normal operation to minimize disk consumption, and increase the amount of detail when debugging.

  2. Click OK to save the configuration.

Create a Profile

You can create a profile that enables you to monitor one or more counters and instances of a Windows NT performance object.

 Follow these steps:

  1. Right-click on the profiles list under the Status tab, and select Add Profile. 
    The Profile properties dialog appears.
  2. Specify a name for the profile and add any additional information for the profile in the Description field.
  3. Under the Object Selection sub-tab, select the performance object to be monitored from the Object drop-down.
  4. Select a counter for which the value is extracted from the Counters drop-down.
  5. Select the instance of the selected counter from the Instance drop-down. When a performance object has several instances (for example Process), a specific instance is identified here. If an object has multiple instances, * is identified, and the watcher is applied to all the instances.

    Notes:

    • When you select an Object and a Counter of a .NET application under monitoring, there can be instances with similar name. Add a new key value in your system registry to show instances in the application_pprocessID format. For more information, see the Known Issues and Workarounds section in the ntperf (Performance Collector) Release Notes.
    • You can monitor the objects and their respective counters and instances by entering an * in the Counter and Instance fields.
    • If you do not select an object, counter and instance from their respective lists, then, it creates an incorrect combination of object, instance and counter. In this case, the probe displays the error message, Invalid combination of object, counter and instance. Please check

    • CA recommends you to create a new profile to monitor the same performance object and instance but a different counter.
  6. Select the Calculate value using scaling ratio checkbox to calculate the value using scaling ratio. This option enables you to adjust the counter scale settings to enhance the visibility of counter data. The counter values can be scaled exponentially from .0000001 to 1000000.
  7. Define the scaling ratio that is used in the Calculate value using scaling ratio field. 
  8. Click OK, and Apply to save the configuration.

Note: Configure at least one of Alarm on Value, Alarm on Value Change, or Miscellaneous sections.

Configure Alarm on Value 

You can configure the alarms on value to monitor the values of the performance object. If this value is exceeded, the specified alarm is generated.

Follow these steps:

  1. Right-click in the profiles list under the Status tab, and select Add Profile.
    The Profile properties window appears.
  2. Update the following information in the Alarm on Value sub-tab:
    • Watch: specifies the threshold parameters to monitor the counter value and generate alarms.
      • Threshold Operator: specifies the threshold operator to compare the actual value with the threshold value.
        Default: Blank

        Note: The probe sends an alarm when the reverse of the defined threshold operator is met. This condition applies to all threshold operators.
        Example: If the condition is defined as >=1 and the actual value is < 1, the probe generates an alarm.

        For the ntperf and ntperf64 probes, the alarms are generated after comparison of the actual value with the threshold value.

        For example, 
        • Example 1
          CitrixMonitor-DB Connected: CitrixMonitor/Database Connected CitrixMonitor//Database Connected = 1.0000 | 1.0000 != (limit= 1.000000) failed
          In this example, the expected value of CitrixMonitor//Database Connected was not equal to one (!= 1). But, the actual value was 1. Since this expectation failed, the alarm was generated.
        • Example 2
          CitrixMonitor DB Trans Erros per Sec: CitrixMonitor/Database Transaction Errors per sec CitrixMonitor//Database Transaction Errors/sec = 0.0000 | 0.0000 > (limit= 0.000000) failed
          In this example, the expected value of CitrixMonitor DB Trans Erros per Sec was expected to be greater than 0 (>0). But, the actual value was 0. 
          Since this expectation failed, the alarm was generated.
      • Threshold Value: specifies the threshold value to generate alarms.
        For example, if your threshold operator is <= (less than or equal to), threshold value is 1250, and the monitor value is 1355; then the probe generates an alarm. If the monitor value is below or equal to 1250, no alarm is generated in the given scenario.
        Default: Blank

        Note: CA recommends you to provide or select a value in the Threshold Operator and Threshold Value fields as the probe does not send alarms if these fields are blank.

      • Fetch Current: retrieves the current value of the counter and displays it in the Threshold Value field. You can increase the Configurator request timeout value on the Setup tab, if the retrieve option fails.
    • Sampling: specifies the number of performance values for which average based on the Check Interval period (specified in the Setup tab) is computed. For example, if the Sampling value is 5 and the Check Interval is 60 seconds, then average of the latest 5 performance values received during each 60-second period is computed.

      Note: For some counters that work on delta (difference of the current and previous value), the sample value must be greater than 1 and need to be configured accordingly. In this scenario, the first value of QoS will be null, as there is no previous value to measure the delta.

      Important! The sample value should not be less than 1.

    • Alarm message: specifies the alarm message to be generated when the threshold value breaches. This value overrides the built-in alarm message definition.
  3. Click OK.

Configure Alarm on Value Change

You can configure the alarm on value change threshold to monitor the change in the measured values of the performance object. 

Follow these steps:

  1. Right click in the profiles list under the Status tab, and select Add Profile.
    The Profile properties window appears.
  2. Update the following values on the Alarm on Value Change sub-tab:
    • Watch: specifies the parameter for tracking the change in the counter values.
      • Threshold Operator: specifies the threshold operator to compare the actual value with the threshold value.
        Default: Blank

        Note: The probe sends an alarm when the reverse of the defined threshold operator is met. This condition applies to all threshold operators. Example: If the condition is defined as >=1 and the actual value is < 1, the probe generates an alarm.

      • Threshold Value: specifies the threshold value to generate alarm.
        For example, if your threshold operator is <= (less than or equal to), threshold value is 100, and the change in monitor value is 105; then the probe generates an alarm. If the change in monitor value is below or equal to 100, no alarm is generated.
        Default: Blank

        Note: CA recommends you to provide or select a value in the Threshold Operator and Threshold Value fields as the probe is unable to send an alarm if these fields are blank.

    • Sampling

      • Change summarized over all samples: Select this option to make a summary of changes in the counter value. For example, you have configured 3 in the Samples field of the Alarm on Value tab; the probe calculates delta QoS between 3rd and 1st sample values and the QoS value as the average of three samples. If the three samples have a value of 10, 12, and 20 respectively; the probe calculates the delta QoS value as 10, which is the difference between 3rd and 1st value (20-10). And the QoS value as 14, the average of the three samples.
      • Change between each sample: Select this option to record changes in each sample individually. For example, you have configured 3 in the Samples field of the Alarm on Value tab; then probe calculates delta QoS between 3rd and 2nd sample values and the QoS value as the average of three samples. If the three samples have a value of 10, 20, and 27 respectively; the probe calculates the delta QoS value as 7, which is the difference between 3rd and 2nd value (27-20). And the QoS value as 19, the average of the three samples.
        Default: Change summarized over all samples

      Notes:

      • If the value in the Samples field on the Alarm on Value tab is set as one, then delta value is always zero. 
      • For some counters that work on delta (difference of the current and previous value), the sample value must be greater than 1 and need to be configured accordingly. In this scenario, the first value of QoS will be null, as there is no previous value to measure the delta.

      • The probe uses the current values of the performance object to calculate the QoS data for change between measurements. 

      Important! The sample value should not be less than 1.

    • Messages
      • Message: specifies the alarm message to be generated when changes are beyond the threshold limit.
        Default: AlarmOnDelta
      • Clear Message: specifies the clear alarm message.
        Default: ClearOnDelta
     
  3. Click OK.

Configure Alarm on Daily and Hourly Change

You can configure one or more of the following metrics to configure the QoS properties of the performance object and set the conditions on threshold values to generate alarms. 

  • Daily Metric: Use Daily Metric to monitor the change in the incoming performance data over the specified number of samples once, each day.
  • Hourly Metric: Use Hourly Metric to monitor the change in the incoming performance data over the specified number of samples during an hour. 

Follow these steps:

  1. Right click in the profiles list under the Status tab, and select Add Profile.
    The Profile properties window appears.
  2. Update the following values on the Miscellaneous sub-tab :
    • Not Found message: specifies a message (text, severity level, and subsystem) to be issued if the object is not found.
    • For the Quality of Services messages on section, enable the value checkboxes for the QoS messages you want to send for this profile:
      • Value: enables you to configure the QoS properties and set the conditions on threshold values for generating alarms.
      • Hourly Change: enables you to to monitor the change in the incoming performance data over the specified number of samples during an hour.
      • Changes between measurement: enables you to monitor the change in the measured values of the performance object.
      • Daily change: enables you to monitor the change in the incoming performance data over the specified number of samples during a day. 
  3. Select the When the object has instances, add the Instance name to the QoS table name checkbox to add the instance name to the QoS table name.
  4. Select the Override QoS Target if you want to use ProfileName-ObjectNameCounterNameInstanceName as the QoS target on the USM. If this option is not selected, ProfileName-InstanceName is used.

    Important! CA does not recommend you to change the QoS target after initial configuration. If you change the target later, multiple graphs are displayed on the Unified Service Management (USM) Metrics view (one for every value). The probe can also generate multiple alarms.

  5. Select the QoS Max Value field to define the maximum value of QoS.
    • Define the maximum value of QoS in the Max Value (user defined) field.
    • Select the counter for the specified QoS max value.
  6. Select the Calculate value using scaling ratio option to specify a scaling ratio to calculate the value.
  7. Click OK.

Create Alarm Messages

The probe maintains a pool of messages for different severity types such as Error Message, Warning Message and so on. You can add or edit messages, as required. 

Follow these steps:

  1. Click the Messages tab.
    The Messages pool list appears. 
  2. Right click in the Messages pool list and select New Message.
  3. Update the following information:
    • Message text: specifies the text to be displayed for the message. You can use variables by typing $ and using the desired variable from the displayed list.
    • Severity: select a severity level for the alarm from the drop-down list.
    • Message Token: select a token for the message from the drop-down list.
    • Message subsystem: specifies a descriptive text name for the subsystem. You can use the following variables in the text field:
      $watcher
      $description
      $operation
      $limit
      $value
      $object
      $instance
      $counter
      $profile
      $hostname
      $robotname
      $samples 
  4. Click OK to close the Message properties dialog.
    The New message dialog appears.
  5. Define appropriate message name and click OK to save the new alarm message.
    A new message is created.

Notes: 

  • To edit existing alarm message details, right-click in the Messages pool list and select Edit Message. The message name appears on the title of Message Properties dialog and is non-editable.
  • The New message dialog does not appear when you edit an existing message.
Was this helpful?

Please log in to post comments.

  1. Shu Qiang Li
    2017-07-19 11:32

    we can put a waring for setting a samples value less than 1 in Alarm on Value or Alarm on Value Change tab. The default or the blank value for parameter samples_value is 1 , therefore it is inappropriate to say 'The Sample value must always be greater than 1.'

    1. Raka Saha
      2017-07-20 01:24

      Thanks Shu Qiang Li, for some counters that work on delta (difference of the current and previous value), the sample value must be greater than 1 and need to be configured accordingly. in this scenario, the first value of QoS will be null, as there is no previous value to measure the delta.

      For the value less than 1, we will update the document at the earliest. Please let us know if you have further questions. 

      -Documentation Team.