Explain spot architecture.

Spot architecture refers to designing cloud workloads to run on spot instances—unused compute capacity sold at a discount. The key idea is to build fault‑tolerant, interruptible systems that can tolerate sudden termination. First, use an autoscaling group with a mix of on‑demand and spot instances, and set a maximum price below which the instances will be launched. Employ spot fleets or instance groups to request multiple instance types and availability zones, increasing the chance of fulfillment. Design stateless services or use external state stores (e.g., DynamoDB, S3, RDS) so that any instance can be replaced without data loss. For batch jobs, use job schedulers that can pause and resume tasks, or split jobs into small, idempotent units. Implement graceful shutdown hooks to checkpoint progress before termination. Use Spot Termination Notices (available 2 minutes before eviction) to trigger cleanup or migration. Finally, monitor spot market prices and adjust the bid strategy or switch to on‑demand when prices spike. This architecture maximizes cost savings while maintaining high availability and resilience.