An ECS service runs on EC2 instances and is fronted by an ALB. The ALB spans two Availability Zones, and the ECS service desired count is 2 tasks. The underlying EC2 capacity uses an Auto Scaling group (ASG) with min size set to 1, and the ASG also spans only one subnet in practice. What is the most effective change to meet the requirement that the service continues during a single-AZ instance loss?

Set the ECS deployment configuration to maximum percent 100 so tasks replace instances faster during rollouts.

Deployment timing can help during releases but does not ensure capacity exists in another AZ during failures.

Enable ALB connection draining longer than expected so existing connections survive longer during an AZ event.

Connection draining affects in-flight sessions but does not restore capacity for new requests after AZ loss.

Reduce task memory reservations to pack both tasks onto a single EC2 instance.

Packing tasks onto fewer instances increases blast radius; it does not improve multi-AZ resilience.

What does this SAA-C03 question test?

CIDR notation defines the prefix length.

What is the correct answer to this question?

The correct answer is: Increase ASG min size to at least 2 and ensure the ASG uses subnets in at least two Availability Zones. — The failure described is loss of instances in one Availability Zone. If the ASG keeps only one instance across one AZ effectively, tasks have nowhere to run when that AZ becomes impaired. Increasing the ASG min size and ensuring it uses subnets across at least two Availability Zones creates redundant compute capacity. With ECS tasks scheduled onto available instances across those AZs and ALB health checks controlling traffic, the service can continue serving requests during a single-AZ event. Why others are wrong: Option A changes rollout behavior but not steady-state placement or capacity redundancy across AZs. Option C may help existing connections but cannot prevent loss of new healthy targets when the AZ fails. Option D increases risk by concentrating tasks on a single instance, which worsens resilience rather than improving it.

What should I do if I get this SAA-C03 question wrong?

Then try more questions from the same exam bank and focus on understanding why the wrong options are tempting.