How we migrated a live routing system using AI-assisted refactoring

AI summary

Datadog engineers refactored their Stream Router control plane by migrating to a new storage architecture while maintaining live production traffic. They leveraged AI tools to accelerate the implementation of method-level refactoring by feeding models existing code, new schemas, and failing tests. The success of the migration relied on a robust test-driven development process to validate AI-generated code against a high-stakes, large-scale metrics pipeline.

Key points

• AI is most effective for refactoring when utilized as a tool for narrow, well-defined tasks rather than architectural decision-making.
• The reliability of AI-assisted code is directly constrained by the quality and coverage of the existing test suite.
• Migrating critical production control planes requires a robust validation strategy, such as blue/green deployments running against live data for extended periods.
• Providing context-rich prompts—including existing implementations, new target schemas, and failing tests—enables LLMs to perform effective code translation.