Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?

 
ON THIS PAGE
 

Observability Recommendation Engine Workflow

Observability Recommendation Engine Overview

Defining and manually instantiating KPI rules for each device in your network can be tedious and error-prone. With real-time changes, your monitoring requirement can change and therefore the KPI rules should also evolve.

The Observability Recommendation Engine feature in Routing Director helps you to move away from the manual, device-by-device KPI setup toward an automated approach. This not only saves time but also reduces human error and improves scalability

Observability Recommendation Engine evaluates device configurations, device inventory attributes (platform, OS, version, and so on), and user-defined preferences, and looks for new configuration changes. Based on the device configurations, Observability Recommendation Engine automatically generates KPI recommendations tailored to each device, and suggests new KPIs or delete the non-relevant KPIs if the specific configurations were deleted from the device. Observability Recommendation Engine introduces a new KPI sub-menu called Smart KPI Assistant. For more information about the new GUI functionality, see About the Smart KPI Assistant Page.

Once the KPI recommendations are instantiated, you can view the collected KPI data on the Troubleshoot Devices page (Observability > Health > Troubleshoot Devices). For more information, see About the Troubleshoot Devices Page.

Tip:

The following are key terms that appear throughout this topic and are explained here for your reference:

  • Greenfield devices—Devices are newly introduced in the network. Once the devices are onboarded, the configurations are done via the network implementation plan.

  • Brownfield devices—Existing devices in the network that are already configured. Routing Director does not automatically discover or monitor these pre-existing configurations.

  • System KPIs—System KPIs are system-generated default KPIs. These KPIs are automatically monitored and analyzed by Routing Director to assess device health. Routing Director collects these KPIs based on predefined rules. For more information, see System Default KPIs in Observability.

  • Custom KPIs—Custom KPIs are user-defined KPIs. You can use Routing Director to define and monitor KPIs tailored to your specific network needs. Routing Director collects these KPIs based on custom rules. For more information, see System Default KPIs in Observability.

  • Rule—Rule is a representation of a KPI. A rule defines what data to pick up from the device, where to look for that data in the device configuration, how to transform it and represent it as a KPI.

  • Recommendation—Recommendation is a suggestion generated by the Observability Recommendation Engine that certain KPIs (rules) are applicable to a particular device.

The following table outlines how device monitoring and KPI instantiations differ with and without the Observability Recommendation Engine:

  Without Observability Recommendation Engine With Observability Recommendation Engine
Greenfield devices

  • Once KPIs are selected based on the network implementation plan, there is no control over the instantiated KPIs. You cannot enable or disable the KPIs, or edit the KPI thresholds.

  • Automatically enables monitoring of relevant system KPIs based on device configurations.

  • Flexibility to enable or disable KPIs for recommendation.
Brownfield devices

  • Only Hardware KPIs are monitored.

  • To monitor other KPIs (interface, routing and mpls, and so on), you had to manually create and instantiate custom KPIs.

  • Automatically enables monitoring of relevant system KPIs based on device configurations.

  • System KPIs can now be instantiated automatically.
System KPIs

  • Only Hardware KPI support for brownfield devices.

  • No flexibility to modify KPI thresholds.

  • Supported across greenfield and brownfield devices.

  • Automatically enabled and monitored.

  • Supports modification of KPI thresholds.
Custom KPIs

  • Required manual creation and instantiation on each brownfield device.

  • Automatically instantiated based on the device configuration.
Tagging core interfaces Core interfaces are tagged with the role:core tag. Core interfaces are tagged with the role:core tag.

Automatic or Manual Approval of KPIs

Observability Routing Engine also provides the flexibility to govern the KPIs that must be instantiated in the following two ways:

  • Automatic approval of KPIs—KPIs are instantiated automatically without requiring manual intervention. You can select automatic approval if you understand what the KPI intends to monitor. Automatic approval is preferred when you do not have any changes to be made to the KPI thresholds.

    Note: For new installations, system KPIs are auto approved by default and you can directly view the collected KPI data on the Troubleshoot Devices page (Observability > Health > Troubleshoot Devices). For more information, see About the Troubleshoot Devices Page.

  • Manual approval of KPIs—You can approve or reject the KPIs based on your specific requirements. You can select manual approval if you want to review the KPI recommendations before you approve them. Manual approval is recommended if you want to modify the KPI thresholds or control which interfaces are monitored.

Threshold Modifications

Default threshold values are defined for system KPIs and custom KPIs. Routing Director allows you to edit the threshold values (also referred to as rule variables in the GUI) as per your requirements. You can edit the threshold values in the following methods:

  • Device level—Use the Recommendation Details page (Observability > Health > Smart KPI Assistant > KPI Recommendations > Detection column > KPI added link) to edit the threshold values at the device level. The device level threshold values takes precedence over the other two methods. For more information, see About the Recommendation Details Page.

    Note: For custom KPIs, you can also edit the threshold values in the Rule Instances tab (Observability > Health > Smart KPI Assistant > KPI Workspace > Rule Instances). For more information, see About the Rule Instances Tab.

  • Organization level—Use the Rules tab (Observability > Health > Smart KPI Assistant > KPI Workspace > Rules) to edit the threshold values at the organization level. The organization level threshold values takes precedence over the threshold values defined at the rule level. For more information, see Edit a Rule Variable.

  • Rule level—Use the Rules tab (Observability > Health > Smart KPI Assistant > KPI Workspace > Rules) to edit the threshold values for a specific rule. You can download the rule, modify the threshold values, and later upload the rule with new threshold values. The rule level threshold values have the least precedence among the three methods. For more information, see Edit a Rule Variable.

Interface Role Detection and Monitoring

Observability Recommendation Engine (whether enabled or disabled) analyzes the device configurations, identifies which interfaces are core facing interfaces (only physical interfaces), and tags these core interfaces with the role:core tag. Observability Recommendation Engine later starts monitoring the core interfaces. For more information, see About the Recommendation Details Page.

To view the list of interfaces that are tagged role:core, see the Interfaces tab (Observability > Troubleshoot Devices > Device-name > Inventory > Interfaces). For more information, see About the Interfaces Tab.

Benefits

  • Reduces the risk of unintended telemetry and processing load by separating rules that can be applied automatically from those that require approval, and by preventing automatic roll out of newly added default rules to existing devices unless you explicitly scope them.

  • Keeps monitoring aligned with the network state by re-evaluating and reconciling recommendations when device configuration, device and interface tags, or the rule catalog changes.

  • Supports a consistent operational approach in mixed brownfield and greenfield environments by providing recommendation-based coverage when intent plans are not available.

Enable Observability Recommendation Engine for System KPIs on Newly Installed Routing Director

Use the following workflow to enable the Observability Recommendation Engine for system KPIs on a newly-installed Routing Director:

Note:

These steps can only be performed by a user with Network Admin or Super User role. Users with Observer role have read-only access and can only view the KPI data.

  1. Log in to the Routing Director GUI. For more information, see Access the Routing Director GUI.

  2. Create a new network implementation plan by clicking the + (Add) icon on the Network implementation Plan page (Inventory > Device Onboarding > Network Implementation Plan).

    The Create Network Implementation Plan page appears.

  3. On the Create Network Implementation Plan page:

    1. Enable the Observability check box.

    2. Click + Add new device profile. The Create Device Profile page appears. On the Analytics tab, enable ORE and Smart KPI Engine toggle buttons.

      Note:

      We recommend you to enable ORE and Smart KPI Engine to make sure that KPI monitoring is aligned with the device intent.

    3. Click save to create a network implementation plan.

    For more information, see Add a Network Implementation Plan.

  4. Upon successful creation of the network implementation plan, on the Network Implementation Plan page, select the network implementation plan and click More > Trigger Onboarding to initiate the onboarding process.

    Upon successful onboarding, Observability Recommendation Engine starts generating KPI recommendations for devices that are onboarded using Routing Director.

    All relevant system KPIs are automatically monitored.

  5. Navigate to the Troubleshoot Devices page (Observability > Health > Troubleshoot Devices) to view the collected KPI data. For more information, see About the Troubleshoot Devices Page.

    Note:

    Even though the system KPIs are automatically enabled, if required, you can disable the system KPIs for recommendation. Use the Rules tab (Observability > Health > Smart KPI Assistant > KPI Workspace > Rules) to disable a system KPI for recommendation. For more information, see About the Rules Tab.

Enable Observability Recommendation Engine for Custom KPIs on Newly Installed Routing Director

Use the following workflow to enable the Observability Recommendation Engine for custom KPIs on a newly-installed Routing Director:

Note:

These steps can only be performed by a user with Network Admin or Super User role. Users with Observer role have read-only access and can only view the KPI data.

  1. Log in to the Routing Director GUI. For more information, see Access the Routing Director GUI.

  2. Create a new network implementation plan by clicking the + (Add) icon on the Network implementation Plan page (Inventory > Device Onboarding > Network Implementation Plan).

    The Create Network Implementation Plan page appears.

  3. On the Create Network Implementation Plan page:

    1. Enable the Observability check box.

    2. Click + Add new device profile. The Create Device Profile page appears. On the Analytics tab, enable ORE and Smart KPI Engine toggle buttons.

      Note:

      We recommend you to enable ORE and Smart KPI Engine to make sure that KPI monitoring is aligned with the device intent.

    3. Click save to create a network implementation plan.

    For more information, see Add a Network Implementation Plan.

  4. Upon successful creation of the network implementation plan, select the network implementation plan on the Network Implementation Plan page and click More > Trigger Onboarding to initiate the onboarding process.

  5. Create a custom rule to generate custom KPIs. For more information, see About the Rules Tab.

  6. All custom KPIs are automatically disabled for recommendation and monitoring, by default. However, you can enable custom KPIs for recommendation, if required.

    Use the Rules tab (Observability > Health > Smart KPI Assistant > KPI Workspace > Rules) to enable a custom KPI for recommendation. For more information, see About the Rules Tab.

  7. Navigate to the Custom KPIs accordion (Observability > Health > Troubleshoot Devices > Device Name) to view the collected custom KPI data. For more information, see About the Troubleshoot Devices Page.

    Note:

    • You can evaluate the KPI recommendations provided by the Observability Recommendation Engine on the KPI Recommendations tab (Observability > Health > Smart KPI Assistant > KPI Recommendations). For more information, see About the KPI Recommendations Tab.

    • You can also decide on whether the custom KPIs must be instantiated automatically without requiring manual intervention. Or, you can give the flexibility to the user to approve or reject recommendations based on the monitoring requirements. For more information, see About the KPI Recommendations Tab.

Enable Observability Recommendation Engine for System KPIs After Upgrading Routing Director

Use the following workflow to enable the Observability Recommendation Engine for system KPIs after you upgrade Routing Director:

Note:

  • After a Routing Director upgrade, the service designs must be updated on the upgraded setup. For more information, see Update Service Designs and Instances After Routing Director Upgrade.

  • These steps can only be performed by a user with Network Admin or Super User role. Users with Observer role have read-only access and can only view the KPI data.

  1. Log in to the Routing Director GUI. For more information, see Access the Routing Director GUI.

  2. On the Network Implementation Plan page (Inventory > Device Onboarding > Network Implementation Plan):

    1. Select an existing network implementation plan and click the Edit (pencil) icon.

      The Edit Network Implementation Plan:Network Implementation Plan Name page appears.

    2. On the Edit Network Implementation Plan:Network Implementation Plan Name page (Inventory > Device Onboarding > Network Implementation Plan > Network Implementation Plan Name), enable the Observability check box, and save the network implementation plan.

    For more information, see Edit a Network Implementation Plan.

  3. On the Device and Interface Profiles page (Inventory > Devices > Device and Interface Profiles):

    1. Select a device profile associated with the network implementation plan and click the Edit (pencil) icon.

      The Edit Device Profile: Device Profile Name page appears.

    2. On the Analytics tab, enable ORE and Smart KPI Engine toggle buttons.

      Note:

      We recommend you to enable ORE and Smart KPI Engine to make sure that KPI monitoring is aligned with the device intent.

    3. Save the device profile.

    For more information, see Edit and Delete a Label or Profile.

  4. On the Network Implementation Plan page, select the modified network implementation plan and click the Publish button to publish the network implementation plan. For more information, see Publish a Network Implementation Plan.

    Upon successful onboarding, Observability Recommendation Engine starts generating KPI recommendations for devices that are onboarded using Routing Director.

    All system KPIs are automatically enabled for recommendation and monitoring.

  5. Navigate to the Troubleshoot Devices page (Observability > Health > Troubleshoot Devices) to view the collected KPI data. For more information, see About the Troubleshoot Devices Page.

    Note:

    Even though the system KPIs are automatically enabled, if required, you can disable the system KPIs for recommendation. Use the Rules tab (Observability > Health > Smart KPI Assistant > KPI Workspace > Rules) to disable a system KPI for recommendation. For more information, see About the Rules Tab.

Enable Observability Recommendation Engine for Custom KPIs After Upgrading Routing Director

Use the following workflow to enable the Observability Recommendation Engine for custom KPIs after you upgrade Routing Director:

Note:

  • After a Routing Director upgrade, the service designs must be updated on the upgraded setup. For more information, see Update Service Designs and Instances After Routing Director Upgrade.

  • These steps can only be performed by a user with Network Admin or Super User role. Users with Observer role have read-only access and can only view the KPI data.

  • Use this workflow if the user wants the Observability Recommendation Engine to handle the custom KPI instantiations,

  1. Log in to the Routing Director GUI. For more information, see Access the Routing Director GUI.

  2. On the Network Implementation Plan page (Inventory > Device Onboarding > Network Implementation Plan):

    1. Select an existing network implementation plan and click the Edit (pencil) icon.

      The Edit Network Implementation Plan:Network Implementation Plan Name page appears.

    2. On the Edit Network Implementation Plan:Network Implementation Plan Name page (Inventory > Device Onboarding > Network Implementation Plan > Network Implementation Plan Name), enable the Observability check box, and save the network implementation plan.

    For more information, see Edit a Network Implementation Plan.

  3. On the Device and Interface Profiles page (Inventory > Devices > Device and Interface Profiles):

    1. Select a device profile associated with the network implementation plan and click the Edit (pencil) icon.

      The Edit Device Profile: Device Profile Name page appears.

    2. On the Analytics tab, enable ORE and Smart KPI Engine toggle buttons.

      Note:

      We recommend you to enable ORE and Smart KPI Engine to make sure that KPI monitoring is aligned with the device intent.

    3. Save the device profile.

    For more information, see Edit and Delete a Label or Profile.

  4. On the Network Implementation Plan page, select the modified network implementation plan and click the Publish button to publish the network implementation plan. For more information, see Publish a Network Implementation Plan.

  5. Create a custom rule to generate custom KPIs. For more information, see About the Rules Tab.

  6. All custom KPIs are automatically disabled for recommendation and monitoring, by default. However, you can enable custom KPIs for recommendation, if required.

    Use the Rules tab (Observability > Health > Smart KPI Assistant > KPI Workspace > Rules) to enable a custom KPI for recommendation. For more information, see About the Rules Tab.

  7. Navigate to the Custom KPIs accordion (Observability > Health > Troubleshoot Devices > Device Name) to view the collected custom KPI data. For more information, see About the Troubleshoot Devices Page.

    Note:

    • You can evaluate the KPI recommendations provided by the Observability Recommendation Engine on the KPI Recommendations tab (Observability > Health > Smart KPI Assistant > KPI Recommendations). For more information, see About the KPI Recommendations Tab.

    • You can also decide on whether the custom KPIs must be instantiated automatically without requiring manual intervention. Or, you can give the flexibility to the user to approve or reject recommendations based on the monitoring requirements. For more information, see About the KPI Recommendations Tab.