Articles — brandur.org

Last week I wrote about leaving Stainless and my intention to work on building my side project River into a small, sustainable business.

Fifteen years ago, when some colleagues and I were building Heroku’s V3 API, we set an ambitious goal: the public API should be powerful enough to run our own dashboard.

Author’s note: This is a longer piece that starts off with exposition into the nature of the N+1 query problem.

Listen/notify in Postgres is an incredible feature that makes itself useful in all kinds of situations.

Go’s JSON library provides the option DisallowUnknownFields which even if not intuitively obvious, is handy for adding a layer of improved DX to web APIs.

Years ago I wrote about my trouble with a job queue in Postgres, in which table bloat caused by long-running queries slowed down the workers’ capacity to lock jobs as they hunted across millions of dead tuples trying to find a live one.

One of Go’s best features is not only that it does parallelism well, but that it’s deeply baked in.

Anyone who’s seen a couple different production database environments is likely familiar with the “soft deletion” pattern – instead of deleting data directly via DELETE statement, tables get an extra deleted_at timestamp and deletion is performed with an…

One of the major revelations for almost every new user to Postgres is that there’s no technical advantage of specifying columns as varchar(n) compared to just using bound-less text.

After a few months of research and experimentation with running a heavily DB-dependent Go app, we’ve arrived at the conclusion that sqlc is the figurative Correct Answer when it comes to using Postgres (and probably other databases too) in Go code beyond…

For years, like most of the rest of the internet, I’ve been using Google Analytics to measure traffic to this site without thinking about it too much.

Big data has an unfortunate tendency to get messy.

A few months ago, I wrote about how SortSupport works in Postgres to vastly speed up sorting on large data types 1 like numeric or text, and varchar.

Many of us in technology-adjacent circles have been grumbling for years about the continued centralization of the web.

For the last couple weeks I’ve been making a few upgrades to the backend that generates this site (previously), with an aim on rebuilding it to be faster, more stable, and more efficient.

Most often, there’s a trade off involved in optimizing software.

You start building your new project.

If you read through the release notes for upcoming Postgres 11, you might see a somewhat inconspicuous addition tucked away at the bottom of the enhancements list: Many other useful performance improvements, including making ALTER TABLE ..

Deep inside a mountain in Texas, a clock is being built.

Aug 3, 2025 update: This article was originally written in 2018.