Mongo: An unexpected journey
summary
Project starts on Postgres then moves to MongoDB and hits slow queries as lists grow from ~5 items to very long. Team adds indexes, uses in-memory cached list updated by a homebrew state machine and Sidekiq jobs to avoid repeated lookups — that fixes the list latency. For analytics they prototype Google Sheets via the Google API (jobs take 30–60 minutes and kill dynos) then adopt a third‑party ETL (staged laser) to sync into a SQL BI store. Front end moves to React and a custom serializer gem handles inconsistent document shapes.
AI-written summary · about
chapters (12)
- 0:00Introduction to journey
- 1:08Overview of MongoDB challenges
- 3:18Starting with MongoDB
- 5:51Optimizing MongoDB performance
- 7:19Legacy data structure issues
- 9:28Caching list of care falls
- 11:15Data accessibility struggles
- 13:26Solving data analytics with Google Sheets
- 14:53Moving data to PostgreSQL
- 17:04Frontend scalability and React
- 19:12Building a custom serializer
- 21:07Reflections and future plans
