The requirements are high availability, resilience to hardware failures, and automatic scaling.

• Option A is correct because using a multi-zonal node pool distributes nodes across multiple zones, ensuring resilience during zone-specific hardware failures and maintaining high availability.
• Option C is correct because combining the Cluster Autoscaler (to adjust nodes) and Horizontal Pod Autoscaler (to adjust pods) ensures the application scales automatically to handle traffic spikes and resource demands.

Other options:

• Option B (multiple zonal node pools) is less efficient than a single multi-zonal pool for even distribution.
• Option D conflicts with GKE's native Cluster Autoscaler, which handles node scaling.
• Option E relies on manual intervention, violating the requirement for automatic scaling.