ON THIS PAGE
 

Add an Analytical Processor

Analytical Processors Overview

Analytical processors receive data from sources or other processors, analyze it, and prepare it for further analysis. Each processor has at least one input and one or more outputs.

Analytical processors fall into the following categories:

  • Grouping

  • Check

  • Periodic

  • Arithmetic

  • Specialized

  • Miscellaneous

Our use case highlights two analytical processors for BFD monitoring: the State processor and the Match Count processor. In the following sections, we'll configure these processors to run simultaneously and apply specific functions to the output of the source processor output.

Add a State Processor

The State Processor analyzes defined conditions and generates Boolean outputs to trigger anomalies. It compares values against a predefined state, receives Discrete State as input, and generates a Boolean output. The output returns true when the input matches the specified criteria, enabling anomalies. In BFD monitoring, the processor monitors the BFD state for each neighbor. The processor triggers an anomaly whenever the state is 'Down', 'Init', 'Init' or 'Failing'.
To add a State processor:
  1. From the left pane in the Apstra GUI, select your blueprint, and navigate to Analytics > Probes.
  2. From the Probes page, select the BFD Monitoring processor and click the Edit button under the Actions column.
  3. Click Add Processor.
    Scroll down to the Analytics Processors category and navigate to Check > State.
  4. Rename the State and Output fields. Because we are monitoring down sessions, we renamed the fields to Down_BFD_Sessions.
  5. Click Add to add the State processor to the pipeline

    Our goal is to monitor the BFD Local State and raise an anomaly if the state is 'Down','Init', or 'Failing'.

  6. Enter the following values:
    In our example, the output of the source processor functions as input for the analytical processor.

    • Inputs:

      • Stage Name: BFD-State

      • Column Name: BFD Local State

    • Raise Anomaly: True. This setting triggers an alert when the comparison result is true.

  7. Click Update Probe.

    The processor outputs true for sessions in the 'Down', 'Init', or 'Failing' state, (see the Value column), and false for all others. The processor detects five anomalies, indicating that five BFD sessions are Down. All sessions are linked to the generic system (to_generic), as shown in the Link column.

    The following example shows a negative (false) value. A false value indicates that all BFD sessions are operational.

Add a Match Count Processor

As a grouping processor, the Match Count Processor requires a Group_By configuration parameter that defines the grouping scheme. The processor produces results based on the defined grouping scheme. In the pipeline, the Match Count processor serves as the final stage.

To add a Match Count processor:
  1. From the Probes page, select the BFD Monitoring processor and click the Edit button under the Actions column.
  2. Click Add Processor.

    Scroll down to the Analytics Processors category and navigate to Grouping > Match Count.

  3. Rename the Match Count and Output stages. We renamed both stages to Down_Session_Count.
  4. Click Add to add the Match Count processor to the pipeline.
  5. Enter the Match Count values:

    For our example, we set the values to the following:

    • Input Stage:

      • Stage Name: Down_BFD_Sessions

      • Column Name: Value

    • Processing:

      • Group by: system_id and Link. The processor receives input from the State processor output (true or false) and counts those values to generate a numeric total.

      • Reference State: true. This setting triggers an alert when the comparison result is true.

  6. Click Update Probe.
You finished adding analytical processors. Next, continue to Create the Dashboards and Widgets to learn how to create dashboards and add widgets to the processors.