Configure QoS Thresholds

NNM iSPI Performance for QA QoS thresholds enables you to track the health and performance of the QoS interfaces and nodes in your network. This includes the following topics:

You can configure the thresholds based on the following QoS element types:

  • QoS Class
  • QoS Node Group
  • ClosedQoS Parent Policy A parent policy contains references to other policies, that are known as child policies. You can define thresholds only on the parent policies. However, NNM iSPI Performance for QA applies the parent policy threshold on the classes configured for the child policies too.
  • Independent QoS Policy (a policy that does not refer to any other policies)

You can establish thresholds for the probes associated with the QoS elements. You can configure these thresholds to create an incident whenever the network performance measurement assigned to the site breaches a threshold.

NNM iSPI Performance for QA performs the following actions if a threshold is breached:

  • Sets the QoS element status to Major.
  • Creates an incident for the violated threshold.
  • Retains the threshold state as Nominal, or sets the threshold state to High or Low depending on the count-based, or time-based threshold configuration.

The global manager receives the threshold states from the sites in the regional managers. The thresholds configured for the QoS elements of the global managers are not applicable for the sites of regional managers.

You can monitor the network performance and generate an incident based on the count-based threshold or time-based threshold configuration. However, you can only configure either a count-based or time-based threshold configuration for a combination of a QoS element and metric.

Threshold Configurations

Count-Based Threshold Configuration

You can generate an incident based on the count or number of consecutive times a metric violates the threshold value. You can define this count in the Threshold Configuration form.

Time-Based Threshold Configuration

Time-Based threshold configuration enables you to raise an alert when the threshold breached state persists for more than a specific time period. This is derived by specifying X as the duration of time in minutes when the metric is in a threshold breached state within Y number of minutes specified in the sliding window.

Example for Time-Based Threshold Configuration

Consider a scenario, where the polling interval is 5 minutes; High duration is 10 minutes; and High Duration Window is 60 minutes. In this scenario, an incident is generated whenever the High Duration exceeds 10 minutes within the 60 minute duration. NNM iSPI Performance for QA uses a sliding window wherein each time the High Duration (10 minutes) is reached, NNM iSPI Performance for QA drops the oldest polled value (first 5 minutes) and adds the most recent (between 60 to 65 minutes). This procedure continues and enables you to determine time-based threshold violation.

You can make utmost use of the Time-Based threshold violation by ensuring that the duration specified in the sliding window is greater than or equal to the polling interval.

Add QoS Threshold Configuration

To add a new QoS threshold:

  1. Launch the QoS Threshold configuration form.

    1. Select Configuration workspace > Quality Assurance Configuration Console. The configuration console opens.
    2. In the Configuration workspace, select Threshold Configuration > QoS. The QoS Threshold Configuration form opens.
  2. In the Configured QoS Thresholds panel of the QoS Threshold Configuration form, click New.
  3. On the Threshold Settings tab, click New to configure the metrics for the threshold.
  4. Use any one of the following options to complete creating the threshold:

    Icon Description
    Close Closes the Add Threshold Configuration form without saving the threshold information you have entered.
    Save and Close Saves the threshold information and closes the Threshold Configuration form.
  5. Click Apply Threshold Now in the QoS Threshold Configuration form to apply the threshold immediately (otherwise, the threshold is applied at the next discovery cycle).

    To view the changes in the QoS Threshold Configuration form, click Refresh.

  6. Check the following log file if you see an error:

    Linux:$NnmDataDir/log/qa/qa.log

    Windows:%NnmDataDir%\log\qa\qa.log

Add QoS Threshold Settings

To configure the metrics for the threshold:

  1. Specify the following to configure the threshold settings:
  2. Field Name Description
    Type Select the type of threshold violation. The valid types are Count-Based and Time-Based.
    Metric

    Select the metric for which you are configuring the threshold.

    After you select the metric to configure the threshold, the list of fields relevant to the selected metric appear. You can specify the following values to configure the new threshold:

    Field Name Description
    High Value

    Enter the threshold value. This value indicates the maximum value above which the metric will be considered to have violated the Nominal range. For Packet Loss metric, enter the High Value in percentage.

    High Value Rearm

    Enter the high value rearm for the threshold. For Packet Loss metric, enter the High Value Rearm in percentage.

    The High Value Rearm is used to indicate the end of the high threshold state and NNM iSPI Performance for QA clears the incident once it reaches below this value.

    The high value rearm must always be lower than the high value.

    Example

    For the Discarded Packets percentage, you must generate an incident when the percentage is 80 and clear the incident when the percentage comes down to 60.

    Set the following values for the threshold:

    • High Value: 80
    • High Value Rearm: 60

    This value enables you to be aware when a network performance problem starts to improve.

    The following field appears, if you selected the Type as Count-Based:

    Field Name Description
    Trigger Count

    Specify after how many consecutive threshold violations NNM iSPI Performance for QA must alert the operator by transitioning the threshold state to High.

    The following fields appear if you selected the Type as Time-Based:

    Field Name Description
    High Duration

    Designate the minimum time within which the metric value must remain in the High range.

    For example if you specify this value to be 20 minutes for Packet Loss Percentage metric, NNMi considers the threshold to be violated if the Packet Loss Percentage is high for 20 minutes.

    You define the high threshold value in the High Value field.

    The High Duration should be equal to or greater than the associated Polling Interval setting, because that is how often NNMi provides a data point..

    High Duration Window

    Designate the window of time within which the High Duration criteria must be met.

    To enable this setting, the value must be:

    • greater than 0 (zero)
    • the same as or greater than the High Duration value

    The NNMi uses a sliding window. Each time the High Window Duration is reached, NNMi drops the oldest polling interval and adds the most recent.

    For example, if you specify this value to be 30 minutes for Packet Loss Percentage metric, NNMi considers the threshold to be violated if the Packet Loss Percentage is high for 20 out of 30 minutes.

    Select the following to generate an incident when the time-based threshold or count-based threshold value is violated:

    Field Name Description
    Generate Incident Select this option if you want NNM iSPI Performance for QA to generate an incident for count-based or time-based threshold violations. By default this option is selected.

  3. Use any one of the following options to complete the task:

    Icons Description
    Close Closes the Add Threshold Settings form without saving the threshold information you have entered.
    Save and Close Saves the threshold information and closes the Threshold Settings form
  4. Continue creating the threshold in the Add QoS Threshold Configuration form.

Edit QoS Threshold Configuration

To edit an existing QoS threshold:

  1. Launch the QoS Threshold configuration form.

    1. Select Configuration workspace > Quality Assurance Configuration Console. The configuration console opens.
    2. In the Configuration workspace, select Threshold Configuration > QoS. The QoS Threshold Configuration form opens.
  2. You can edit the following settings:
  3. Field Name Description
    Order

    Specify a numeric value. NNM iSPI Performance for QA checks for configuration settings in the order you define (lowest number first). NNM iSPI Performance for QA uses the first match found for each threshold. Provide a unique ordering number for each threshold.

    Thresholds with duplicate Order numbers are checked in random order.

    Policy

    Specify a QoS policy name on which you want to configure the threshold and click to add the policy in the list.

    The QoS elements on which the selected policy is applied come under the threshold.

    Class

    Specify a QoS class name on which you want to configure the threshold and click to add the class in the list.

    The QoS elements on which the selected class is applied come under the threshold.

    Node Group Specify a QoS node group on which you want to configure the threshold and click to add the node group in the list.

    Make sure that you have specified at least one criterion for the threshold. That is, specify at least one policy, class, or node group for the threshold.

    If you create a new threshold configuration or modify the threshold configuration criteria (policy, class, or node group), NNM iSPI Performance for QA applies the changes in the next configuration polling cycle. However, NNM iSPI Performance for QA does not delete the incidents that are already generated for an existing threshold.


  4. Use any one of the following options to complete modifying the threshold:

    Icons Description
    Close Closes the Add Threshold Configuration form without saving the threshold information you have entered.
    Save and Close Saves the threshold information and closes the Threshold Configuration form
  5. Click Apply Threshold Now in the QoS Threshold Configuration form to apply the threshold immediately (otherwise, the threshold is applied at the next discovery cycle).

    To view the changes in the QoS Threshold Configuration form, click Refresh

Check the following log file if you see an error:

Linux:$NnmDataDir/log/qa/qa.log

Windows:%NnmDataDir%\log\qa\qa.log

Edit QoS Threshold Settings

To configure the metrics for the threshold:

  1. Make sure that you have specified the mandatory fields in the editing QoS threshold configuration.
  2. Select the threshold settings, and Click Edit
  3. Specify the following to configure the threshold settings:
  4. Field Name Description
    Type Select the type of threshold violation. The valid types are Count-Based and Time-Based.
    Metric

    Select the metric for which you are configuring the threshold. The metrics are populated based on the service.

    After you select the metric to configure the threshold, the list of fields relevant to the selected metric appear. You can specify the following values to configure the new threshold:

    Field Name Description
    High Value

    Enter the threshold value. This value indicates the maximum value above which the metric will be considered to have violated the Nominal range. For Packet Loss metric, enter the High Value in percentage.

    High Value Rearm

    Enter the high value rearm for the threshold. For Packet Loss metric, enter the High Value Rearm in percentage.

    The High Value Rearm is used to indicate the end of the high threshold state and NNM iSPI Performance for QA clears the incident once it reaches below this value.

    The high value rearm must always be lower than the high value.

    Example

    For the Discarded Packets percentage, you must generate an incident when the percentage is 90 and clear the incident when the percentage comes down to 60.

    Set the following values for the threshold:

    • High Value: 90
    • High Value Rearm: 60

    This value enables you to be aware when a network performance problem starts to improve.

    The following field appears, if you selected the Type as Count-Based:

    Field Name Description
    Trigger Count

    Specify after how many consecutive threshold violations NNM iSPI Performance for QA must alert the operator by transitioning the threshold state to High.

    The following fields appear if you selected the Type as Time-Based:

    Field Name Description
    High Duration

    Designate the minimum time within which the metric value must remain in the High range.

    For example if you specify this value to be 20 minutes for Packet Loss Percentage metric, NNMi considers the threshold to be violated if the Packet Loss Percentage is high for 20 minutes.

    You define the high threshold value in the High Value field.

    The High Duration should be equal to or greater than the associated Polling Interval setting, because that is how often NNMi provides a data point..

    High Duration Window

    Designate the window of time within which the High Duration criteria must be met.

    To enable this setting, the value must be:

    • greater than 0 (zero)
    • the same as or greater than the High Duration value

    The NNMi uses a sliding window. Each time the High Window Duration is reached, NNMi drops the oldest polling interval and adds the most recent.

    For example, if you specify this value to be 30 minutes for Packet Loss Percentage metric, NNMi considers the threshold to be violated if the Packet Loss Percentage is high for 20 out of 30 minutes.

    Select the following to generate an incident when the time-based threshold or count-based threshold value is violated:

    Field Name Description
    Generate Incident Select this option if you want NNM iSPI Performance for QA to generate an incident for count-based or time-based threshold violations. By default this option is selected.

  5. Use any one of the following options to complete the task:

    Icons Description
    Close Closes the Add Threshold Settings form without saving the threshold information you have entered.
    Save and Close Saves the threshold information and closes the Threshold Settings form
  6. Continue modifying the threshold in the Edit QoS Threshold Configuration form.

If you modify the threshold settings or update the monitored metrics, NNM iSPI Performance for QA applies the changes in the next polling cycle. For example, You have a threshold T1 that monitors the metric Dropped Packets. If you changed the configured threshold value for the metric from 5 to 10, NNM iSPI Performance for QA applies the changes in the next polling cycle.

However, NNM iSPI Performance for QA does not delete the incidents that are already generated for an existing threshold. For example, if an incident was already generated for threshold T1, NNM iSPI Performance for QA does not delete the incident when the metric value is changed from 5 to 10.

For a list of incidents generated for NNM iSPI Performance for QA threshold violations, see Threshold Incidents

Delete QoS Thresholds

To delete an existing QoS threshold:

  1. Launch the QoS Threshold configuration form.

    1. Select Configuration workspace > Quality Assurance Configuration Console. The configuration console opens.
    2. In the Configuration workspace, select Threshold Configuration > QoS. The QoS Threshold Configuration form opens.
  2. Select a threshold in the Threshold Settings panel and click Delete.
    or
    Click Delete All to delete all QoS thresholds.
  3. Click Refresh in the Configured QoS Thresholds panel to view the changes.

However, NNMi does not delete the incidents that are already generated for an existing threshold.

Import QoS Thresholds

To import threshold configurations from an XML file:

  1. Launch the QoS Threshold configuration form.

    1. Select Configuration workspace > Quality Assurance Configuration Console. The configuration console opens.
    2. In the Configuration workspace, select Threshold Configuration > QoS. The QoS Threshold Configuration form opens.
  2. Click Import.
  3. In the user prompt dialog, enter the file name from where you want to import the QoS threshold configuration information.

    You must enter the file name with full path information; for example, C:\temp\CBQoSthreshold_conf.xml

  4. Click OK in the user prompt dialog.

    If a threshold is already defined and displayed in the QoS Threshold Configuration panel, the import utility does not import the configuration information for this threshold from the XML file.

You can also import threshold configuration information using the following command line utility:

Linux: $NnmInstallDir/bin/nmsqathresholdconfigutil.ovpl –u <username> –p <password> –import –type cbqos <filename>

Windows:%NnmInstallDir%\bin\nmsqathresholdconfigutil.ovpl –u <username> –p <password> –import –type cbqos <filename>

If the threshold import fails, check the following log files:

Linux: $NnmDataDir/log/qa/qa.log

Windows:%NnmDataDir%\log\qa\qa.log

-u <username> -p <password> are optional parameters.

Export QoS Thresholds

To export the existing threshold configurations to an XML file:

  1. Launch the QoS Threshold configuration form.

    1. Select Configuration workspace > Quality Assurance Configuration Console. The configuration console opens.
    2. In the Configuration workspace, select Threshold Configuration > QoS. The QoS Threshold Configuration form opens.
  2. Click Export.
  3. Type the file name where you want to export the existing QoS threshold configuration in the user prompt dialog.

    You must type the file name with full path information; for example, C:\temp\CBQoSthreshold_conf.xml

    If you type the XML file name without entering the absolute path, by default the file gets saved in the following directory:

    Linux: $NnmDataDir/shared/qa/conf

    Windows : %NnmDataDir%\shared\qa\conf

  4. Click OK in the user prompt dialog.

You can also export the existing QoS threshold configuration using the following command line utility:

Linux: $NnmInstallDir/bin/nmsqathresholdconfigutil.ovpl –u <username> –p <password> –export –type cbqos <filename>

Windows:%NnmInstallDir%\bin\nmsqathresholdconfigutil.ovpl –u <username> –p <password> –export –type cbqos <filename>

The threshold export utility does not export a threshold unless the threshold is associated with at least one site.

If the threshold export fails, check the following log files:

Linux:$NnmDataDir/log/qa/qa.log

Windows:%NnmDataDir%\log\qa\qa.log

-u <username> -p <password> are optional parameters.

Supported QoS Threshold Configuration Metrics

You can configure threshold on the following metrics based on the selected device type:

QoS Metrics & Juniper CoS Metrics

Metric Description
Pre Policy Bit Rate (kbps) The bit rate of the traffic per class before applying the CBQoS policy, measured in kbps
Post Policy Bit Rate (kbps) The bit rate of the traffic per class after applying the CBQoS policy, measured in kbps
Packet Drop (%)

Percentage of the packets dropped per class.

It is calculated using the following formula:

(Total number of dropped packets / total number of packets transmitted per class)*100

Exceeded Packets (%)

Percentage of the packets dropped per class due to exceeded policies

It is calculated using the following formula:

(Total number of packets dropped due to exceeded policies / total number of packets transmitted)*100

Violated Packets (%)

Percentage of the packets dropped per class due to violated policies

It is calculated using the following formula:

(Total number of packets dropped due to violated policies / total number of packets transmitted)*100

Discarded Packets (%)

Percentage of the packets dropped per class due to the queuing action

It is calculated using the following formula:

(Total number of packets dropped due to the queuing action / total number of packets transmitted)*100

Queue Utilization (%)

Utilization rate for the queue

It is calculated using the following formula:

(Queue depth/Maximum queue depth) * 100

Queue Bandwidth Utilization (%)

Percentage of the bandwidth utilized per class

Available only when the bandwidth reservation per class is measured as one of the following values:

* As absolute value

* As a percentage of the total bandwidth. It is calculated using the following formula:

(PostPolicyBytes in kbps / Bandwidth configured in kbps) * 100

Dropped Shape Packets (%)

Percentage of packets dropped per class due to the shaping action

It is calculated using the following formula:

(Total number of packets dropped due to the shaping action / Number of packets transmitted for the selected class) * 100

Delayed Shape Packets (%)

Percentage of packets delayed per class due to the shaping action.

It is calculated using the following formula:

(Total number of packets delayed due to the shaping action/total number of packets transmitted)*100

RED Packets Tail Drop (%)

Percentage of packets dropped per class due to greater number of packets in the queue than the maximum threshold

It is calculated using the following formula:

(Total number of packets dropped by the RED algorithm / total number of packets transmitted)*100

RED Packets Drop (%)

Percentage of packets dropped per class due to the buffer overflow

It is calculated using the following formula:

(Total number of packets dropped by the RED algorithm / total number of packets transmitted)*100

Marked DSCP Packets (%)

Percentage of packets marked with IP DSCP bits per class

The class sets a configured DSCP value for the incoming IP packets.

It is calculated using the following formula:

(Packets with the IP DSCP bit set / total number of packets transmitted) * 100

Marked IP Precedence Packets (%)

Percentage of packets marked with IP Precedence per class

The class sets a configured Precedence value for the incoming IP packets.

It is calculated using the following formula:

(Packets with the IP precedence bit set / total number of packets transmitted) * 100

Marked FRDE Packets (%)

Percentage of packets marked with IP FRDE bits per class

The class sets a configured FRDE value for the incoming IP packets.

It is calculated using the following formula:

(Packets with the IP FRDE bit set / total number of packets transmitted) * 100

Pre Policy Bit Rate (kbps)

Bit rate of the traffic before applying the Tricolor policy, measured in kbps.

It is calculated using the following formula:

(jnxCosQstatQedByteRate * 8) / 1000

Post Policy Bit Rate (kbps)

Bit rate of the traffic after applying the Tricolor policy, measured in kbps.

It is calculated using the following formula:

(Current transmit rate of the queue * 8) / 1000

Packet Drop (%)

Percentage of the packets dropped per class.

It is calculated using the following formula:

(Total dropped packets / total number of packets queued at the output queue) * 100

RED Packets Drop (%)

Percentage of packets dropped per class due to the buffer overflow

It is calculated using the following formula:

(Total number of packets dropped by the RED algorithm / total number of packets queued at the output queue) * 100

RED Packets Tail Drop (%)

Percentage of packets dropped per class due to greater number of packets in the queue than the maximum threshold

It is calculated using the following formula:

(Total number of tail dropped packets / total number of packets queued at the output queue * 100

RED LP Packets Drop (%)

Total number of low priority packets dropped per class after applying the Tricolor policy, expressed as a percentage of the total packets dropped.

It is calculated using the following formula:

(Total number of low priority packets dropped / total dropped packets ) * 100

RED HP Packets Drop (%)

Total number of high priority packets dropped per class after applying the Tricolor policy, expressed as a percentage of the total packets dropped.

It is calculated using the following formula:

(Total number of high priority packets dropped / total dropped packets ) * 100