Spot instances

Spot instances are virtual machines that a cloud provider sells from its unused capacity at a large discount compared with on-demand pricing. The trade-off is that the provider can reclaim them when it needs the capacity back. AWS calls them EC2 Spot Instances, while Google Cloud and Azure call them Spot VMs.

How it works

On AWS, spot prices can be up to 90% lower than on-demand prices for the same instance type. When AWS reclaims capacity, it sends an interruption notice two minutes before stopping or terminating the instance. Your workload must drain and shut down within that window. Spreading across several instance types and Availability Zones lowers the chance that one capacity shortage hits every node at once.

Why it matters in production

Spot capacity can cut compute cost sharply for stateless and fault-tolerant workloads. It is risky for single-replica services, local data and anything that needs a guaranteed replacement quickly. In Kubernetes, a reclaim can evict pods without honouring a PodDisruptionBudget, so teams pair spot nodes with multiple replicas, an interruption handler and an on-demand fallback.