What are the Route 53 routing policies and when would you use each?
Simple returns a record. Weighted splits traffic by percentage. Latency sends users to the lowest-latency Region. Failover switches to a backup when the primary is unhealthy. Geolocation routes by user location. Multivalue returns several healthy IPs.
Explain it simply
Route 53 is the call-centre operator deciding which branch office answers your call. She can give everyone the same number (simple), send 10% of callers to the new office (weighted), connect you to the closest office (latency), or switch to another office when the main one is shut (failover).
Example
Simple one record, same answer for everyone
Weighted 70% -> old stack, 30% -> new stack (canary releases)
Latency user in Mumbai -> ap-south-1, user in London -> eu-west-2
Failover primary healthy -> primary; unhealthy -> secondary (needs health check)
Geolocation users in India -> Indian-language site, EU users -> EU site
(compliance, content localisation)
Geoproximity route by distance to resources, with adjustable bias
Multivalue up to 8 healthy records returned, simple client-side balancingLatency routing is based on measured network latency, not on physical distance. Geolocation is based on where the request comes from and is used for legal or content rules, so always add a default record for locations that do not match.
Health checks work with failover, weighted, latency and multivalue policies to remove unhealthy endpoints from answers.
What to say in the interview
"Weighted for gradual rollouts, latency for the fastest Region, failover for active-passive disaster recovery with health checks, and geolocation for content or legal rules by country. Simple is for a single resource."
Common mistakes
- Using geolocation when the real goal is lowest latency.