For compute workload architecture, which load balancer and instance type configuration meets Mountkirk's requirements?

The correct option uses a global load balancer with managed instance groups and autoscaling policies, along with non-preemptible Compute Engine instances. A global load balancer distributes traffic across multiple regions, providing high availability and disaster recovery. Managed instance groups (MIGs) automate instance creation, deletion, and updates, ensuring consistent configurations. Autoscaling policies dynamically adjust the number of instances based on demand, optimizing cost and performance. Non-preemptible instances are suitable for workloads requiring continuous availability and predictable performance, as they are not subject to preemption by Google Cloud. Network load balancers operate at Layer 4 and are regional, offering less flexibility and global distribution compared to global HTTP(S) load balancers. Preemptible instances are cost-effective but can be terminated at any time, making them unsuitable for Mountkirk's workload, which requires high availability and uninterrupted service.