You are designing a highly available web application on Compute Engine. The application must survive a zone failure without manual intervention. How should you configure the Compute Engine instances?

Deploy the instances in a Zonal Managed Instance Group and configure a Cloud Monitoring alert to manually recreate them in another zone if the primary zone fails.

Deploy the instances in a Regional Managed Instance Group (MIG) across multiple zones.

Deploy the instances in an Unmanaged Instance Group and attach it to a Global Load Balancer.

Use a single Compute Engine instance with a Regional Persistent Disk.