Custom Telemetry Collection Overview

Juniper Apstra Custom Telemetry Collection allows you to define new telemetry services for data monitoring and analysis in Apstra. It also allows you to customize analytics according to your business needs. This process simplifies the previously complex task of adding a telemetry service, which required advanced programming and knowledge of the telemetry SDK.

With the custom telemetry collection, you can do the following:

Run the Junos CLI show commands that provide you with the data you want analyzed.
Identify the specific key and value to extract from the show command based on its XML output.
Create a telemetry collector definition.

Here are some examples of what you can do with the custom telemetry collection. Any show command on a switch can serve as an example.

Monitor various counters (firewall filter match count, IRB interface statistics, and so on).
Monitor device power usage and thermal output for capacity planning.
Monitor device health (line card status or other environmental statuses).
Monitor protocol status or features enabled with configlets (BFD, MACsec, QoS, multicast, OSPF, RPM and so on).
Monitor power usage and thermal output for capacity planning.

This document outlines a use case for monitoring BFD Protocols using a custom telemetry collector.

Use Case

The use case monitors the operational status of BFD sessions across all network layers where administrators enable BFD. This includes fabric BGP sessions at both the underlay and overlay levels, as well as any external BGP peerings. While Juniper Apstra offers built-in BGP monitoring, Data Center administrators can gain additional insights by monitoring BFD.

Because BFD functions as a lightweight hello protocol with much shorter, configurable timers compared to BGP keepalives, it helps administrators detect forwarding-path failures faster. A useful way to distinguish between the two: a BFD anomaly indicates a forwarding path or liveness issue, while a BGP anomaly points more to a control-plane routing or session problem, so the two protocols complement each other well.

In this guide, we'll demonstrate the complete workflow, including data ingestion, data processing, and data visualization. In this walkthrough, we'll monitor multiple BFD session types as an example.