[Proposal] Extract Sec codebase from monolith

Our team often loses time and confidence when unrelated flaky tests in the monolith break our CI runs. Even small, isolated changes require running the full test suite, which is slow and costly. To reduce this friction, I propose that we extract our domain code into a dedicated Rails engine. Instead of depending directly on the monolith’s models, the engine would replicate only the data it needs through an event or similar solution. This would allow us to own our domain end-to-end, with fast and reliable CI that runs only our specs, while still integrating with the rest of the system through stable, event-driven boundaries.

This approach gives us three benefits: speed (we test independently), reliability (we aren’t blocked by unrelated flakes), and clarity (a smaller, focused codebase for our developers to navigate). At the same time, it preserves integration with the larger platform by consuming a controlled feed of core data from the monolith, rather than reaching into its internals. We’d start with a lightweight replication mechanism, gradually move our reads fully into the engine’s database, and keep the monolith as the system of record. This sets us up for long-term scalability and cleaner boundaries, without forcing a big-bang rewrite or service split.