Driving Architectural Improvements with Dependency Metrics

This 28-minute Linux Foundation talk explores Spotify's journey of migrating client codebases to a Bazel-powered monorepo while focusing on reducing build times. Learn how Spotify developed specialized data processing pipelines, metrics, and dashboards to analyze their massive dependency graph containing thousands of targets. Discover the Bazel Dependency Metrics system that combines static graph analysis, build data, and developer behavior patterns to identify architectural bottlenecks, assess code changes that negatively impact the organization, and prioritize high-value refactoring targets. The presentation walks through the creation of these metrics, data collection methods, dashboard implementation, and shares early results showing how this approach has guided Spotify's complex migration efforts.