These questions are independently created for study purposes. They are not official AWS exam questions, actual exam questions, leaked questions, or exam dumps.
SAA-038
For a web application running in the Tokyo Region, you want to switch automatically to a standby environment in the Osaka Region when a Region-wide failure occurs. The domain name that users access cannot change. Which configuration is the most appropriate?
Choices
Explanation and answer
Read the explanation and correct answer (spoiler — try answering first)
Correct answer: C. Use a Route 53 failover routing policy with Tokyo as primary and Osaka as secondary, and associate a health check with the primary record
Explanation
To fail over automatically between Regions while keeping the same domain name, combine a Route 53 failover routing policy with a health check. When the primary health check fails, DNS responses switch automatically to the secondary endpoint.
Per-choice explanations
A. Create an Auto Scaling group that spans Tokyo and Osaka so that instances launch in Osaka when Tokyo becomes unhealthy
Incorrect. An Auto Scaling group can span multiple Availability Zones within a single Region, but it cannot span Regions.
B. Use Route 53 weighted routing with weight 100 for Tokyo and weight 0 for Osaka, and swap the weights manually when a failure is detected
Incorrect. Weighted routing itself is a valid feature, but because changing the weights is manual it does not meet the requirement to switch automatically. The time from detection to switchover also depends on how fast people respond.
C. Use a Route 53 failover routing policy with Tokyo as primary and Osaka as secondary, and associate a health check with the primary record
Correct. You can tune the switchover speed with the health check failure threshold and interval, and keeping the DNS record TTL short makes the change take effect quickly.
D. Register EC2 instances from both Tokyo and Osaka in an Application Load Balancer target group and distribute traffic using health checks
Incorrect. The Application Load Balancer itself is a resource inside the Tokyo Region, so a Region-wide failure in Tokyo takes down the ALB as well and no cross-Region switchover happens. Registering instances in another Region as targets is actually possible with an IP address type target group and inter-Region VPC peering (instance ID type targets are limited to the load balancer's own VPC), but the ALB still depends on a single Region.
Practical Use
Amazon Route 53 routing policies are separated by purpose. Simple routing just returns a single destination; weighted routing distributes traffic in a specified ratio (used for canary releases and phased migration); latency-based routing returns the Region with the lowest latency from the client's perspective; geolocation routing returns a destination based on the requester's geography; and failover routing defines primary and secondary paths and returns the secondary only when the primary is unhealthy. Failover routing is the one that corresponds to an active/passive disaster recovery design. The key to automatic switchover is the Route 53 health check. Health checkers around the world send periodic requests to the target endpoint (an IP address or domain name), and by default they mark it Unhealthy when the failure count exceeds the threshold at a 30-second interval (10 seconds for fast checks). In addition, you can base a health check on the state of a CloudWatch alarm, or create a calculated health check that combines multiple health checks with AND / OR. Because the switchover manifests as a change in the DNS response, the actual speed depends on the health check detection time plus the record TTL plus the client-side resolver cache. Keeping the TTL around 60 seconds is the standard practice, and if you need faster switchover, consider a mechanism that does not depend on DNS, such as AWS Global Accelerator. On SAA-C03, requirements such as "automatic failover to another Region during a Regional failure" and "switch over while keeping the same domain name" make failover routing plus a health check the correct answer. Distractors include using an ALB or an Auto Scaling group across Regions, both of which you can eliminate by knowing that they are Regional services. Weighted routing is ruled out by reading that it is manual and therefore does not fit the requirement.