benchmarking – Page 2 – Türker Yıldırım

A build with no bundler

Deep diveJuly 31, 20259 min read

A site with 18 KB of JavaScript and a build pipeline sized for something larger. What a bundler provides, listed, and which of it is needed.

Two years of query logs, read once

Deep diveJune 18, 202510 min read

A slow query log rotating for two years and never aggregated, and performance work driven by whoever complained loudest.

The endpoint that was slow for one customer

Deep diveJune 4, 20259 min read

A healthy p95 and one customer for whom the endpoint takes four seconds. An aggregate metric that averages away the only case that matters.

18 KB of JavaScript, and where the rest went

NoteApril 3, 20252 min read

A page that shipped 140 KB in 2024 and 18 KB in 2025, accounted for honestly. The largest saving is the widget nobody could convert, deferred rather than…

PHPUnit 12, and the tests we deleted instead

Deep diveMarch 13, 20259 min read

A two-hour upgrade, and a directory of 188 tests excluded from every tool since 2023 and running on every push.

Acceptance rate, edit distance, and what neither measures

NoteFebruary 25, 20252 min read

Two metrics that are available, neither of which answers the question anybody is asking. A high acceptance rate is compatible with a tool that produces plausible code and…

Load testing before a launch, again, and the same finding

NoteJanuary 8, 20251 min read

A second launch, a second load test, and the same class of finding as last year for a different reason. Two years, two resource limits, both of them…

A query that got faster after the upgrade, and why

NoteNovember 19, 20241 min read

A query that improved by a factor of twenty after a major version upgrade, examined rather than accepted. Understanding why something got faster is worth the ten minutes…

PHPUnit 11, and a suite that had already been migrated

Deep diveFebruary 29, 20249 min read

11.0 in February, a year after the 10.0 migration did the hard part, and eleven deprecations that were all things 10 had warned about.

The bundle we split by route, and the waterfall it created

Deep diveOctober 24, 202310 min read

A 900 KB bundle, split by route, and a page that got slower. Six sequential requests before the first render.