Schedule API

The original Fastify service had grown organically and carried a Cartesian product in event lookups that made response times unpredictable under load.

Rewrote it in NestJS with one other developer and redesigned the data model: 17 tables in Postgres, managed through Drizzle. The Cartesian product went away via a CTE pre-filter that narrows the result set before joins — a targeted fix with a large impact on query plans.

Redis backs two features: shareable timetable links with configurable TTL, and per-user rate limiting on link generation. Runs in Docker, CI on GitHub Actions.