Behind the Scenes: Block 450 JVM Repositories Into Monorepo to Reduce Dependency Drift

2026-06-19 · Source: InfoQ · Field: Technology & Digital — Software Development & Engineering, Cloud Computing & IT Infrastructure · Depth: Advanced, medium

Summary

Block, Inc. successfully migrated approximately 450 JVM-based repositories from a polyrepo architecture to a monorepo across its Cash App and Square engineering organizations. This large-scale initiative addressed growing coordination and dependency management challenges, including version drift and runtime incompatibilities, inherent in the previous model. The consolidated codebase now supports about 8,800 builds per week, achieving p90 CI times of around 10 minutes on a stable main branch. Key implementation strategies included developing a custom IntelliJ plugin to load only necessary projects, adopting merge queues for high commit volumes, and investing in shared Gradle plugins, dependency graph-based build scoping, and git performance tuning. This transition aimed to simplify cross-service development, enhance dependency visibility, and reduce operational friction.

Key takeaway

For MLOps Engineers or Software Engineers considering a monorepo migration, understand that this architecture demands substantial, ongoing platform team investment. If you cannot commit to properly funding a dedicated platform team, a polyrepo might be a more suitable choice to avoid punishing poorly behaved sibling projects. Ensure your projects are similarly shaped in terms of languages and frameworks to maximize benefits and justify the significant migration effort.

Key insights

Consolidating JVM polyrepos into a monorepo reduces dependency drift and improves developer experience through atomic updates.

Principles

• Monorepos require significant platform team investment.
• Standardized projects benefit most from monorepos.
• Dependency graphs enable efficient build scoping.

Method

Block implemented a custom IntelliJ plugin for project loading, used merge queues, and leveraged shared Gradle plugins with dependency graph-based build scoping and git tuning.

In practice

• Use custom IDE plugins for large monorepos.
• Implement merge queues for high commit volume.
• Define project ownership with tooling.

Topics

• Monorepo Migration
• Dependency Management
• JVM Repositories
• CI/CD Optimization
• Gradle Plugins
• Developer Experience

Code references

• block/gradle-monorepo-ij-plugin
• joshfriend/spotlight
• block/artifact-swap

Best for: CTO, VP of Engineering/Data, Software Engineer, DevOps Engineer, MLOps Engineer

Related on AIssential

• H2O MLOps Enterprise Model Registry & Hugging Face | Part 8 Integration | Part 8 — A robust model repository is essential for managing and governing a large, diverse portfolio of AI models in production.
• Article: Building Production-Ready tRPC APIs: The TypeScript Alternative to Apollo Federation — tRPC offers superior performance and type safety for TypeScript monorepos compared to GraphQL Federation.
• Migrating Your GitHub CI to Hugging Face Jobs — Migrating GitHub CI to Hugging Face Jobs cuts CPU test times by 30% and enables cost-effective GPU testing via custom runners and hardware.
• Enabling Evolutionary Database Development: database branching with Lakebase, continued — Copy-on-write database branching makes per-instance database development feasible, removing key constraints for evolutionary design.
• 6 Docker Tricks to Simplify Your Data Science Reproducibility — Achieve data science reproducibility by meticulously freezing Docker environments at every layer prone to drift.
• Incentives and Evidence in Learned Service Orchestration — Current RL orchestration research often lacks production relevance due to flawed evaluation and institutional publication incentives.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by InfoQ.