All issues
Ship Notes · SN #03·MAY 30, 2026
Boring infrastructure, fast teams
Two notes this week on why the least exciting parts of your stack are often where the biggest wins hide.
1. Pick the database you already know
The novel datastore rarely pays off early. Postgres handles relational data, JSON, full-text search and queues well enough to carry most products to real scale. Reach for something exotic only when a concrete limit forces your hand — not because a benchmark looked good.
2. Logs are a feature, not an afterthought
Structured logs with a request ID turn a two-hour incident into a two-minute one. Add them before you need them. The day you're debugging a payment failure in production is the wrong day to discover you log nothing useful.
Previous issue
SN #02 — Saying no to scope