You need a cost-optimized GKE cluster for an app where some components tolerate downtime and others must remain highly available. How should you configure node pools and schedule workloads?

The correct approach is to use both Spot VMs and standard VMs. Spot VMs offer significant cost savings but can be preempted, making them suitable for fault-tolerant workloads that can handle interruptions. Standard VMs provide consistent availability and are necessary for critical workloads that require high uptime. Deploying critical workloads on the standard node pool ensures their continuous operation, while fault-tolerant workloads on the Spot VM node pool optimize costs without impacting essential services. The other options are incorrect because they either don't leverage the cost savings of Spot VMs effectively or place critical workloads at risk of preemption.