Question

A SysOps administrator manages an Application Load Balancer fronting an Auto Scaling group of EC2 instances that serve long-running file uploads averaging 90 seconds. During scale-in events and deployments, users report that in-progress uploads fail with dropped connections. The default deregistration delay is currently in effect. What is the MOST appropriate action to prevent these failed requests?

Accepted Answer

Increase the target group deregistration delay (connection draining) to a value greater than the longest expected request duration, such as 120 seconds.

Answer

Enable cross-zone load balancing on the Application Load Balancer to distribute connections more evenly.

Answer

Reduce the health check interval and unhealthy threshold so failing targets are removed faster.

Answer

Change the load balancer idle timeout to 1 second so stale connections are closed immediately.