itop data modeling review, a hotfix write-up, and some theory reading
Reviewed a staging-to-main PR on itop, dug into MongoDB best practices versus Postgres for the project, documented a hotfix, and read up on Paxos and ACID.
A mixed day: one real architecture call, a small infra write-up, and a bit of fundamentals reading.
itop.com.br
Reviewed a pull request where the diff looked off: code already on main was showing up as a staging commit being merged back. Walked through the branch and merge flow to confirm whether that was correct or a sign of a tangled history.
The bigger thread was data modeling. The project runs MongoDB in Docker today (Mongo container plus a frontend container on a VPS, managed with Dokploy). I checked the current schema against MongoDB's own data-modeling best practices, then asked the harder question: should this project stay on MongoDB or move to Postgres? Weighed the trade-offs, including whether Supabase is worth it and how connection handling differs between the two. No migration yet, just a clearer picture of what each choice costs.
.hermes
Wrote up a hotfix: how the problem was found and how it was fixed, saved as a short note so the next person (or me) does not have to rediscover it. Infra and pipeline only.
Fundamentals
Read a bit of theory on the side: the Paxos consensus algorithm and what ACID actually guarantees. Useful background for the database decisions above.
Some days the win is not shipping code but knowing which database you should be shipping it to.