This routing policy is fairly simply it is just an active-passive failover. You have one record that is the primary instance, a health check is sent to that and if that instance fails it moves to the secondary record which you can attach an optional health check to. So for example if you have 1.2.3.4 set as primary and 5.6.7.8 set as secondary then 1.2.3.4 will be used until it fails in that case it would switch the value to 5.6.7.8. This policy is useful if you want to ensure your site remains up despite failures.