Amazon Route 53
AWS's DNS service with a 100% availability SLA — where routing policies and health checks turn name resolution into a load-balancing and failover layer.
What is Route 53?
Route 53 is AWS’s managed DNS: it hosts your zones, answers queries from a global anycast fleet, and registers domains. Two features elevate it beyond commodity DNS. Alias records map zone apexes (example.com, where CNAMEs are forbidden) directly to AWS resources — ALBs, CloudFront, S3 — resolving to current IPs with no lookup charge. Routing policies make DNS itself a traffic-steering layer.
Routing policies in practice
Weighted routing splits traffic by percentage — the poor man’s canary and the standard blue/green switch. Latency-based sends users to the lowest-latency region among your deployments. Failover pairs a primary record with a health check and flips to a standby on failure. Geolocation/geoproximity pin jurisdictions to regions (data-residency compliance via DNS). Combined with health checks — Route 53 probes endpoints globally and withdraws unhealthy records — you get DNS-level failover measured in tens of seconds… bounded below by one hard constraint: TTL. Failover cannot beat the TTL cached in resolvers worldwide, which is why failover-critical records run TTLs of 30–60 s, trading query volume for agility.
The details that page people at 3 a.m.
Health checks see your endpoint from the public internet — private-only resources need calculated or metric-based checks. Alias-to-ALB failover depends on the ALB’s own health; layering both without understanding which triggers first produces confusing flaps. And DNS failover is eventually consistent by nature: some resolvers ignore low TTLs, so a fraction of traffic lingers on the failed side — design backends to reject gracefully rather than assuming instant cutover.
What people get wrong
- 24-hour TTLs on records you might need to move — set low TTLs before the migration, days ahead.
- Weighted routing without session affinity thinking: users bounce between versions on every resolution.
- Treating the 100% SLA as “DNS can’t be my outage” — misconfiguration isn’t covered by anyone’s SLA.
Primary source: Amazon Route 53 documentation
Historical figures and technical concepts for informational purposes only. Not technical, professional, legal, or financial advice. Sources: Official Documentation.