Spot VM (GCP)
A Compute Engine VM using Google's spare, unused capacity at a steep discount, with the trade-off that Google can reclaim the instance with short notice. Spot VMs are appropriate only for fault-tolerant, interruption-tolerant workloads like batch processing, never for customer-facing services needing reliable availability.
Frequently Asked Questions
How much cheaper is a Spot VM than a standard on-demand Compute Engine instance?
Spot VMs (Google's successor to preemptible VMs) typically run 60-91% cheaper than on-demand pricing depending on machine type and region, since you're bidding on Google's genuinely unused capacity. The catch is Google can reclaim the instance with as little as 30 seconds' notice via a preemption signal, and there's no guaranteed maximum runtime — unlike the old preemptible VMs' hard 24-hour cap, Spot VMs can theoretically run indefinitely if capacity stays available.
What's the right way to architect a workload around Spot VM interruptions?
Use them behind a Managed Instance Group with autoscaling so Google's reclaim events are automatically backfilled, and design jobs to checkpoint progress frequently so a reclaimed instance loses minimal work — batch processing, CI runners, and stateless rendering are ideal. Never put a stateful, single-instance database or a customer-facing API with no redundancy on a Spot VM; the 30-second termination notice usually isn't enough time to gracefully drain live user traffic.