A network operations team wants a small Python service that polls SD-WAN Manager every few minutes and raises a ticket when it sees an edge router's control connections drop, while leaving all device configuration untouched. Which SD-WAN Manager API family should the service query for that live control status?

The monitoring APIs are the right choice because the service needs to inspect each edge router’s current control-connection state and health on a recurring schedule. Reading that operational status is a non-configuring action, so it supports detection without changing device configuration. The alarms APIs are a reasonable incident-integration source: they can expose generated alerts, including alerts associated with a control-connection failure. However, an alarm represents a condition detected and reported by the system; it is not a reliable substitute for querying the live state of every connection, particularly when the service must determine current status rather than depend on alert generation or lifecycle. The events APIs provide timestamped changes useful for reconstructing what happened, but event records describe transitions, not necessarily the router’s present control-connection state. The statistics APIs offer useful recent interface or transport measurements for performance and troubleshooting, but those metrics do not directly report the required control-connection status and health. Exam caveat: API family names and endpoint details can vary by SD-WAN Manager release, and alarms or events may complement monitoring rather than replace it. Practical check: in the target release’s API documentation or a read-only test, confirm that the selected monitoring endpoint returns per-edge control-connection state and health fields, and verify the service account has only the required read permissions.