A company runs a stateful application on EC2 instances behind an Application Load Balancer in an Auto Scaling group. During scale-out events, new instances require about 4 minutes to download configuration files, warm a local cache, and register with an internal service registry before they can serve traffic. Currently, the ALB begins routing requests to instances as soon as they pass the EC2 status checks, causing intermittent errors from instances that are not yet fully initialized. Which approach best ensures instances only receive traffic after initialization completes?