We have identified the root cause of the downtime. High latency on a primary uplink was causing repeated failovers between the primary and secondary uplink. Redirecting all traffic over to the secondary link has resolved the availability problems.

Monitoring is alerting to KenyonWX being unavailable.