Question 29
Domain 5: Incident and Event ResponseA company hosts its staging website using an Amazon EC2 instance backed with Amazon EBS storage. The company wants to recover quickly with minimal data losses in the event of network connectivity issues or power failures on the EC2 instance. Which solution will meet these requirements?
Correct answer: C
Explanation
Amazon EC2 instance recovery is designed for "network connectivity issues or power failures" and works with EBS-backed instances. An Amazon CloudWatch alarm on the StatusCheckFailed System metric can trigger the EC2 "recover" action, which moves the instance to new hardware while preserving the instance’s EBS volumes, minimizing data loss and downtime.
Why each option is right or wrong
A. Add the instance to an EC2 Auto Scaling group with the minimum, maximum, and desired capacity set to 1.
Auto Scaling replaces instances for capacity management, not host recovery while preserving the same instance state.
B. Add the instance to an EC2 Auto Scaling group with a lifecycle hook to detach the EBS volume when the EC2 instance shuts down or terminates.
Lifecycle hooks coordinate instance transitions, but detaching EBS does not provide automatic host-level recovery.
C. Create an Amazon CloudWatch alarm for the StatusCheckFailed System metric and select the EC2 action to recover the instance.
Amazon EC2 instance recovery is the documented remediation for an EBS-backed instance experiencing underlying host impairment such as network connectivity loss or power failure. Under the EC2/CloudWatch integration, a CloudWatch alarm on the StatusCheckFailed_System metric can invoke the EC2 recover action, which restarts the instance on new hardware while retaining the attached EBS volumes; AWS specifically notes this is intended for system status check failures and preserves data on EBS, minimizing both downtime and data loss.
D. Create an Amazon CloudWatch alarm for the StatusCheckFailed Instance metric and select the EC2 action to reboot the instance.
Instance status checks point to guest OS or software issues; reboot is not the right response to host failure.