Route 53 Routing and Health
Explore AWS Route 53 routing and health check mechanisms to understand how DNS routing decisions differ from load balancer target selection. Learn to apply routing policies like failover, weighted, latency-based, and geolocation to meet availability and traffic management goals. Understand troubleshooting steps to diagnose DNS and origin health issues for network reliability.
DNS and a load balancer solve two different problems, and confusing them is where most routing incidents actually go wrong. Route 53 decides which endpoint a client attempts first, by choosing a record set according to a routing policy. An ALB or NLB decides which backend actually serves the request once a connection reaches that endpoint. A healthy DNS answer says nothing about whether the infrastructure behind it can actually handle the request.
Here's a case where that boundary matters concretely. A single hostname needs to serve users globally, fail over between two AWS regions during an outage, and support a small canary rollout, all without changing any client configuration. Those three requirements force two separate decisions, because DNS can only avoid known bad endpoints if a health check marks them unhealthy, while a healthy DNS answer can still lead straight to a 5xx if the load balancer behind it has no healthy targets.
What DNS can and cannot decide
Routing policy selection is about which question Route 53 will answer under pressure, not about how the application behaves once a connection lands. Route 53 decides record answers at query time. An ALB or NLB decides target selection at request time. Those are two different moments, and mixing them up is what makes DNS-layer incidents look like application-layer ones, or the reverse.
Each routing policy optimizes a different dimension, ...