Second latency probe to a backup EPP host

The login screen shows latency to the primary EPP endpoint only. If it also probed a secondary I could pick the healthier route before a big drop.