what’s the difference between Route 53 failover policy and multi value policy
4 Answers
From the AWS documentation,
"Failover routing policy – Use when you want to configure active-passive failover"
"Multivalue answer routing policy – Use when you want Route 53 to respond to DNS queries with up to eight healthy records selected at random"
Imagine a scenario where you had servers in California, USA and North Virginia, USA, but all of your customers are in N. Virgina. Using Failover Routing you can have all of your traffic go to N. Virgina when it’s online and only to California when it’s not. If you were to use the same servers in Multivalue Answer Routing then you would effectively see a 50/50 split of traffic going to both. Your customers going to California may see a degraded user experience.
https://docs.aws.amazon.com/Route53/latest/DeveloperGuide/routing-policy.html
Basically, Multivalue policy is just a simple routing policy that uses health checks to automatically failover
On the other hand, Failover policy does its failover in a manual way where you have to set your active(primary destination) and passive (backup destination) for failovers
From my understanding, in short, Failover Policy puts the management of the "failure re-routing" on AWS (as long as you have configured the health checks), while Multivalue Answer Policy puts the management of the "failure re-routing" on client applications (no health check required).
Long story short would be:
Multivalue does a Round Robin around healthy targets, which means that it will randomly access any healthy target and you have no control over it [as it does not have any weight]. So, if you define 4 records, and 3 are healthy, the client will access any of those 3, randomly without any definition being made to select one of those 3. That is why it is considered a real basic load balancing capability on Route53’s behalf. You can have up to 8 multivalue records.
Failover, only allows you to set a primary and a secondary. Will go always to the Primary unless it goes down. So, if the client, anywhere at anytime, tries to access some kind of resource, it will access only the Primary record unless it is in an unhealthy state, which will rely it to the secondary one. You can only have 2.