SAA-C03 - Design Resilient Architectures (26% of the exam) - Section 2.7

Design fault-tolerant compute and storage using load balancer health checks, redundancy and durable storage.

Describe how Elastic Load Balancing health checks detect unhealthy targets and automatically redirect traffic to healthy instances, and explain why stateless application design simplifies horizontal scaling and fault tolerance. Recognise that Amazon S3 is designed for 99.999999999% durability through redundant storage across multiple devices, and use EBS snapshots to provide point-in-time recovery for block storage volumes.

Elastic Load Balancing health checksAmazon S3 durabilityEBS snapshotsStateless design

Practice question for this objective

Free sampleDesign Resilient Architecturesmedium

An Application Load Balancer fronts an Auto Scaling group, and the team wants unhealthy instances to be replaced automatically rather than just removed from rotation. They configure the Auto Scaling group health check type so the group acts on the load balancer's view of each instance. Which setting makes the Auto Scaling group replace instances that the load balancer reports as unhealthy?

  • AKeep the default EC2 health check type, since it already reflects whether the application is responding to load balancer requests.
  • BSet the Auto Scaling group health check type to ELB so it uses the load balancer target health to decide which instances to replace. Correct
  • CEnable connection draining on the load balancer so in-flight requests finish before the group replaces a failing instance automatically.
  • DLower the load balancer health check interval so failing instances are detected sooner and the group reacts to them more quickly.
An Auto Scaling group must use the ELB health check type to replace instances the load balancer reports as unhealthy. By default an Auto Scaling group only checks EC2 instance and system status. Changing the health check type to ELB makes the group incorporate the load balancer's target health, so an instance that fails the application-level health check is marked unhealthy by the group, terminated, and replaced to maintain the desired healthy capacity.

Why A is wrong: The default EC2 health check only covers instance and system status, so an instance failing application health checks at the load balancer is still considered healthy and is not replaced.

Why B is correct: Setting the health check type to ELB makes the Auto Scaling group treat targets the load balancer marks unhealthy as failed, so it terminates and replaces them automatically.

Why C is wrong: Connection draining controls how existing connections complete during deregistration, but it does not change which instances the Auto Scaling group decides to replace.

Why D is wrong: A shorter interval speeds detection at the load balancer, but unless the group uses ELB health, it never acts on that status to replace the instance.

See more SAA-C03 practice questions, answers explained.

Exam traps in Design Resilient Architectures

Answers that look right on this material and are not. Each one is a distractor from a different question in the SAA-C03 bank for this domain.

  • Enable Auto Scaling group EC2 status checks so the group terminates the instance when its system or instance status check fails on the host.

    Why it is wrong: EC2 status checks verify host and hypervisor reachability, not application responses, so a running instance returning 503 stays in service and keeps receiving traffic.

  • Convert the single EBS volume to a larger Provisioned IOPS SSD volume so it sustains more device failures before any data is lost.

    Why it is wrong: A larger or faster EBS volume still lives in one Availability Zone and a higher IOPS class does not add cross-AZ durability, so a zone failure still loses the data.

  • Attach a General Purpose SSD (gp3) volume to both instances, since any EBS volume can be mounted by multiple instances once the file system supports it.

    Why it is wrong: gp3 does not support Multi-Attach, and a standard EBS volume attaches to only one instance at a time regardless of the file system, so this attach simply fails.

Examworthy is not affiliated with or endorsed by Amazon Web Services. Original, blueprint-aligned practice material only.