Moving Baselime from AWS to Cloudflare: simpler architecture, improved performance, over 80% lower cloud costs

The document outlines Baselime’s migration from an AWS-based data pipeline (Lambda, Kinesis, EC2, ClickHouse, etc.) to Cloudflare’s serverless platform (Workers, Durable Objects, Queues, D1, Workers KV, and Workers Analytics Engine), emphasizing an incremental “strangler fig” approach. The shift reduced operational complexity, improved latency and scalability via global deployment, and cut estimated compute and analytics costs by over 80%, while enabling high-cardinality analytics, built-in rate limiting, and event-driven processing without managing infrastructure.