A team supporting 40 branch routers keeps getting user reports of brief latency spikes, but by the time an engineer logs into each device and runs the usual commands the problem has passed. They also want the polling load on the links to stay light. Which telemetry approach should they adopt to catch these issues while they are happening?
Select an answer to reveal the explanation.
Short Explanation
Use streaming telemetry to send selected measurements to a collector at a configured interval, rather than waiting for someone to poll each router. It’s like getting a doorbell alert instead of checking the porch every hour, helping you catch brief spikes while keeping polling load light.
Full Explanation
Model-driven telemetry lets a router publish selected YANG-modeled data to a collector, either periodically at a configured cadence or, where supported, when specified changes occur. This can reduce reliance on polling and provide measurements frequently enough to help investigate intermittent issues; the available data and cadence depend on the platform, software, and subscription configuration. SNMP polling every five minutes may miss a brief event between polls. An hourly CLI collection cannot provide timely visibility into a short-lived event. A SPAN session can support live packet analysis, but analyzing stored traces after an incident, as described, does not provide timely detection. Telemetry can help narrow the search but does not, by itself, identify the root cause; use it with configuration and syslog context.