A fault-tolerant nightly batch job uses many VMs and costs are too high. You can tolerate VM terminations. What should you do?

The correct answer is to use Spot N2 Standard VMs. Spot VMs are significantly cheaper than regular VMs because they can be preempted (terminated) by Google Cloud if resources are needed elsewhere. Since the job is fault-tolerant and can tolerate VM terminations, Spot VMs are an ideal cost-saving solution. Testing with simulated maintenance events ensures the fault tolerance mechanism works as expected before fully committing to Spot VMs. Using regular N2 Standard VMs (without Spot) would not provide the cost savings needed. Using a Managed Instance Group (MIG) is good for managing groups of VMs, but it doesn't inherently reduce costs for individual VMs; you'd still want Spot VMs within the MIG if cost is the primary concern. Switching to N1 Standard VMs might offer some cost reduction compared to N2, but Spot N2 VMs would likely be more cost-effective and provide better performance than N1, especially for a batch job.