Question

A company runs a stateful web application behind an Application Load Balancer with an EC2 Auto Scaling group. Users report that during scale-in events (when the ASG terminates instances), some active sessions receive HTTP 5xx errors and in-flight requests are abruptly cut off. The operations team wants to ensure instances finish serving existing requests before termination completes, without keeping instances registered indefinitely. Which action BEST addresses this?

Accepted Answer

Configure the target group's deregistration delay (connection draining) to allow in-flight requests to complete before the instance is deregistered

Answer

Increase the ALB health check interval so instances are marked healthy longer during termination

Answer

Enable cross-zone load balancing on the ALB to evenly distribute connections during scale-in

Answer

Switch the ASG termination policy to OldestInstance so long-running sessions are preserved