A staging website runs on a single Amazon EC2 instance backed by EBS. The company wants fast recovery and minimal data loss if the instance experiences network connectivity problems or a host power failure. Which option meets these requirements?

The correct option is to create a CloudWatch alarm for StatusCheckFailedSystem with the EC2 recover action. This directly addresses host-level issues like network connectivity problems or power failures by migrating the instance to healthy hardware while preserving its IP address, instance ID, and all attached volumes (including the EBS volume). This ensures fast recovery and minimal data loss. Adding the instance to an Auto Scaling group with capacity set to 1 doesn't provide recovery for the existing instance in case of a system failure; it would only launch a new instance if the old one terminated, which isn't the primary goal for fast recovery of an existing instance with its data. A lifecycle hook for EBS detachment is irrelevant here as recovery preserves attached volumes. Using StatusCheckFailedInstance monitors issues within the instance, not host-level problems, and a reboot action might not resolve underlying hardware issues, nor does it guarantee data integrity or IP preservation in the same way recovery does.