Explain Spot vs Reserved vs On-Demand Instances.

Example

A startup runs its web API on reserved instances, adds on-demand during traffic spikes, and executes nightly data ETL on spot instances.

Want to master trade-offs like these for interviews?

Explore Grokking System Design Fundamentals, Grokking the Coding Interview, and sharpen your skills with Mock Interviews with ex-FAANG engineers.

Why Is It Important

Picking the right mix can reduce costs by 40–70% while ensuring flexibility, reliability, and scalability.

Interview Tips

• Start with clear definitions.
• Suggest a blended strategy (reserved baseline + on-demand spikes + spot for batch).
• Mention interruption handling with retries, queues, or checkpointing.

Trade-offs

• On-Demand: Flexibility but higher cost.
• Reserved: Big savings but lock-in.
• Spot: Lowest price but risk of termination.

Pitfalls

• Running critical stateful apps on spot.
• Overcommitting to reserved beyond baseline needs.
• Ignoring autoscaling, budgets, or regional differences.