A NOC team is building a dashboard on the SD-WAN Manager monitoring APIs. One tile must show whether each WAN tunnel is up right now and refresh every five seconds, while a second tile plots average latency across the last 30 days for a chosen site. Which split between real-time and historical endpoints should the dashboard use?

The correct split matches each tile to the data it needs: a real-time endpoint supplies current WAN tunnel state for the five-second refresh, while a statistics endpoint supports an average-latency trend over a defined 30-day interval for the selected site. Using a seven-day rolling range is a sound way to display a recent latency trend, and it preserves the correct endpoint types and site scope, but it does not cover the specifically requested 30 days. Polling statistics every five seconds may keep a tile visually fresh, and a 30-day site-scoped statistics query is appropriate for the trend; however, the latest available statistics sample is not a real-time tunnel-state response, so it cannot meet the live-state requirement merely by being polled frequently. Omitting the site filter still produces a valid 30-day latency trend, but it aggregates beyond the chosen site and therefore answers a different question. Exam caveat: endpoint names and available fields vary by SD-WAN Manager version, so distinguish real-time operational state from periodically collected statistics rather than relying on labels alone. Practical check: inspect the request parameters and response timestamps, then verify that the live tile changes within the expected refresh interval and that the trend query returns only the selected site's data across the full 30 days.