Double Buffering with Cloudflare Queues Turns Telemetry Into a Cheap Serverless Data Lake

A telemetry pipeline built on Cloudflare Workers, Queues, Workflows, and R2 uses a double-buffered queue to sort events into UTC-hour batches. A producer routes each payload to an even- or odd-hour queue by modulo math, while an hourly consumer task drains the inactive queue, compresses the batch, and writes it to R2 under a year/month/day/hour path. The result is a partitioned, serverless data lake with very low storage costs.