You are designing a highly available architecture for a mission-critical application on Compute Engine. The application must survive the failure of a single data center without any downtime.

Which TWO architectural decisions should you make to ensure this level of availability? (Select TWO)

To protect against the failure of a single data center, you must deploy your resources across multiple zones (since a zone typically corresponds to a single data center or a collection of closely linked data centers). The best practice for achieving this with Compute Engine is to use a Regional Managed Instance Group (MIG). A Regional MIG automatically distributes your VM instances across multiple zones within a region, ensuring high availability and providing auto-healing capabilities.