Google Automates Spanner Migration with Antigravity CLI, Accelerating Refactoring of 30 DAOs

As the Google Finance Engineering team adopted Cloud Spanner to modernize their legacy data layer, they built an automation pipeline based on Antigravity CLI to enable zero-downtime migration for production services. Previously, manually rewriting dual-write logic across dozens of Data Access Objects (DAOs) was slow and error-prone; this automation solution significantly accelerated migration speed.

Applying a Standardized MutationConverter Pattern

The core of the migration was standardizing the MutationConverter interface to map complex domain models to the Spanner schema. By isolating a dedicated converter class for each DAO to clarify the contract with the Spanner SDK, they established target specifications that allowed AI coding agents to reason and generate code reliably.

Leveraging Antigravity CLI Headless Mode

Instead of using the IDE's interactive chat, they used Antigravity CLI's headless mode (-p) to perform systematic multi-file updates across the entire codebase. Orchestration scripts take target DAOs as input and pass them to Antigravity, then run blaze tests on the generated code. If tests fail, error logs are fed back to enable self-correction.

Large-Scale Overnight Runs and Results

This pipeline supports unattended overnight runs: engineers queue 10 DAOs at the end of the day, and by morning, clean changelists that have been tested and verified are generated. This significantly reduced time spent on repetitive boilerplate refactoring, allowing engineers to focus on high-value tasks such as data modeling and performance optimization.