The database we reach for first
Relational, dull in the best sense, and portable. Most products need a database rather than a platform, and this is the one we would still choose in five years.
Where Postgres earns its place.
Rules in the database
Constraints the application cannot forget, enforced where the data actually lives.
Migrations under review
Schema changes travel in the same pull request as the code that needs them.
Query plans you can read
EXPLAIN before it hurts, not after a customer tells you.
Runs anywhere
Managed, on your own server, or on a laptop: the same engine every time.
Why this, and when not.
Why we choose it
It is the least surprising database there is, and nothing about it is proprietary. You can take it elsewhere on any day you like.
When we would not
A cache, a queue, or a pile of documents with no relationships. A schema you never query relationally is overhead.
Services built on Postgres.
Building something data-heavy?
A 45-minute technical call with the engineer who would lead the build. You leave with a scope sketch either way.