A company runs a legacy system on a single Amazon EC2 instance. The application code cannot be changed, and it cannot run on more than one instance. A solutions architect must design a resilient solution to improve recovery time. What should the architect recommend?

The most effective solution for improving recovery time for a single, unchangeable EC2 instance is to use an Amazon CloudWatch alarm with an EC2 Auto Recovery action. This automatically recovers the instance on new underlying hardware if it becomes impaired due to an underlying hardware issue, without changing its private IP address, EBS volume attachments, or other configuration. Termination protection prevents accidental termination but doesn't address instance failures. Multi-AZ deployment is not possible for a single instance that cannot run on more than one instance. EBS RAID configurations provide storage redundancy but do not address the availability of the EC2 instance itself.