SOA-C02 Practice Question 3: ASG ELB Health Check Type
AWS Certified SysOps Administrator – Associate (SOA-C02) · Test 1 · question 3 of 10
An Auto Scaling group of EC2 instances sits behind an Application Load Balancer. Users report intermittent HTTP 502 errors. Investigation shows that the application process on two instances has crashed: the ALB target group marks those targets as unhealthy and stops routing to them, but the Auto Scaling group leaves them running for hours and never launches replacements. Both instances continue to pass their EC2 status checks.
Which action resolves this?
Show answer and explanation
Answer:
An Auto Scaling group always uses the built-in Amazon EC2 status checks. Other health check sources, including Elastic Load Balancing, are optional and must be turned on for the group. Until the ELB health check is added, a load balancer's unhealthy verdict is invisible to Auto Scaling: the instance is still running and still reachable, so EC2 status checks pass and the instance is never replaced. Turning on the ELB health check makes Auto Scaling honour the target group's verdict and terminate and replace the instance.
Option A treats the symptom. The broken instances stay in the group forever, capacity cost rises, and the same failure recurs on the new instances.
Option C makes the problem worse. A TCP health check only confirms that something is listening on the port; an HTTP check against an application path is far more likely to notice a crashed or hung application, not less.
Option D is wrong and harmful. The grace period only delays the *first* health evaluation after launch, which has nothing to do with instances that have been running for hours. Setting it to 0 can cause new instances to be terminated before the application finishes starting.
Reference: About the health checks for your Auto Scaling group