Skip to content
Refine

Glossary · Pricing & Commitments

What is a Spot Instance?

A spare EC2 capacity instance priced at up to ~90% off on-demand, with the trade-off that AWS can reclaim it with 2 minutes' notice.

By HabileLabs

Definition

Spot Instance

A Spot Instance is an EC2 instance running on AWS's spare capacity, available at a steep discount — typically 60–90% off the on-demand rate. The trade-off is that AWS can interrupt and reclaim the instance whenever spare capacity tightens, with just a 2-minute warning.

Spot is ideal for stateless, fault-tolerant, or batch workloads: CI runners, ML training, data processing, render farms, web tiers fronted by an auto-scaling group with on-demand failover. It is unsuitable for stateful databases or long-running jobs that can't checkpoint.

Why it matters for AWS cost

The cost savings are large enough that many teams should be using Spot for at least part of their EC2 fleet. The reason most teams don't is the perceived complexity of handling interruptions — but modern tooling (Spot fleets, capacity-optimized allocation, ASG mixed-instances policies) handles that automatically for stateless workloads.

Common gotchas
  • !Capacity-optimized allocation is almost always better than lowest-price — fewer interruptions for a few cents more.
  • !Spot interruption rates spike during regional capacity squeezes — design for graceful degradation.
  • !Spot is not eligible for Savings Plans / RI discounts; you're trading commitment savings for interruption risk.

Find waste like this in your own AWS account free

60-second IAM-role connection. Read-only. No signup needed for the bill analyzer.

Refine is built and supported by HabileLabs, an AWS Advanced Tier Services Partner.

Back to glossary15 terms in the AWS cost & FinOps glossary