Let’s DO this: detecting Workers Builds errors across 1 million Durable Objects

Cloudflare describes building an automated error detection system for Workers Builds that parses and normalizes build logs (e.g., Wrangler errors) to categorize failures, using Workers Queues (BuildErrorsQueue), Durable Objects (BuildBuddy and a BuildErrorsAgent with alarms), KV, Hyperdrive/Postgres, and concurrent consumers to backfill results across more than one million builds. The insights enabled proactive fixes—clearer error messages, better package manager handling, bun.lock support, Python runtime selection, and improved caching in monorepos—resulting in faster iteration and higher build completion rates.