The best choice is to modify the Auto Scaling group to use three instances across each of two Availability Zones. This configuration ensures that if one Availability Zone experiences an outage, the remaining Zone can still serve traffic. The Application Load Balancer automatically distributes incoming traffic across healthy targets in multiple Availability Zones, allowing the system to remain available even during failures in one zone. This meets the high availability requirement while preserving the existing application configuration.

Why the other options are incorrect:

Creating an Auto Scaling group across two Regions is not valid because Auto Scaling groups cannot span AWS Regions. Multi-Region architectures require much more complexity, such as global DNS routing and data replication, which is unnecessary for solving the current high availability requirement.

Creating an Auto Scaling template for launching instances in another Region is not a high availability solutionit only helps with manual recovery or disaster recovery. It does not ensure uptime during failures or automate failover operations.

Changing the ALB to use round-robin configuration is a misunderstanding. The ALB already distributes traffic across targets, and configuring round-robin does not address the Availability Zone limitation. The problem is not with traffic distribution but with all resources being isolated to a single zone.

Why this question matters: Understanding how to architect applications across multiple Availability Zones is essential for building fault-tolerant systems on AWS. This question evaluates your ability to use existing services like Auto Scaling and ALB in a resilient and AWS-recommended manner.

Exam tip: If a question involves improving availability and resilience without reworking the application, look for solutions that distribute resources across Availability Zones. Avoid options involving Regions unless the scenario explicitly requires cross-Region recovery.

Full AWS Practitioner Certification Question