Help us improve your experience.

Let us know what you think.

Do you have time for a two-minute survey?



Analytics Overview

Managed devices generate large amounts of data over time. On their own these data are voluminous and unhelpful. With Intent-Based Analytics (IBA) you can combine intent from the graph with current and historic data from devices to reason about the network at-large.

Data generated by devices are ingested via agents and sent to the Apstra server. With the use of IBA probes, data can be aggregated across devices in response to operator configuration. Combining probes with intent from the blueprint graph generates a reduced set of data that can be more easily reasoned about. You can directly inspect advanced data from the Apstra GUI or from REST API to gain real-time insight about the network. It can also be streamed out with our existing streaming infrastructure. Also, based on the state of this advanced data, anomalies can be raised.

While operating IBA at scale, using many probes, disk usage can grow significantly within the Apstra server VM. This is expected because the system will persist at least enough samples to maintain data for the requested duration for all time-series for all existing probes. Additionally, the system will create checkpoint (backup) files up to a configured limit. Settings in the /etc/aos/aos.conf file indicate how often to rotate logs and remove old checkpoint files. Using IBA can increase disk usage to tens of gigabytes. If this is an issue, you can adjust the log rotation settings to reduce disk usage.

Additional space may be used by system snapshots and old images from any in-place Apstra server upgrades. These can be deleted or moved off the system to increase free disk space.

Analytics Dashboard

Analytics dashboards monitor the network and raise alerts to anomalies. Specific dashboards are automatically created and enabled based on the state of the active (operational) blueprint. You can also instantiate predefined dashboards and create your own.

Some other characteristics of analytics dashboards include:

  • You cannot configure the trigger logic that determines when dashboards are auto-created, but you can create/instantiate your own dashboards.
  • Probes that you've created and not modified are reused instead of creating duplicates of those probes.
  • Widgets within each dashboard monitor different aspects of the network and raise alerts to relevant anomalies.
  • When you enable a dashboard, the required probes and widgets are instantiated. If you update or delete associated probes and/or widgets, the dashboard may enter an invalid state. Invalid dashboards are not automatically repaired.
  • You can display analytics dashboards on the blueprint dashboard to have additional network information on one screen. To add them, turn ON the analytics dashboards' default toggles.
  • When upgrading the controller, the auto-creation behavior of dashboards occurs on preexisting active blueprints, in the same way as for newly-created blueprints.

From the blueprint, navigate to Analytics > Dashboards to go to the analytics dashboard. You can create, clone, edit, and delete analytics dashboards. System-generated dashboards are labeled with System and user-generated (and user-modified) dashboards are labeled with the user's name. Dashboards can be shown in various levels of detail by choosing a Display mode (summary, preview, expanded).