Engineering Blog
Long-form write-ups of real engineering work: database scaling, PHP performance, microservices, observability, GitOps and design systems.
Latest article
Microservices split your data across many databases, and everyday work needs it back together. A tour of Lizard: screens built from your schema, records with everything related, an AI doctor that finds missing relationships, queries fixed by AI, dashboards from a sentence, and why it's all safe.
Request objects threaded through every layer, central route wiring, middleware whose scope depends on line order, and framework-specific HTTP types. An author's account of each problem and the design response.
Push-based deployments leave production defined by its history. Here is what changed when six hundred services moved to a pull-based model with Git as the source of truth.
Why a database can run out of connections while its CPU sits idle, why PHP and Knex make that hard to fix in code, and how a protocol-aware proxy resolves it.
PHP-FPM rebuilds your application on every request. What it costs, what changes when the application stays in memory, and the bugs you must remove first.
Incidents are slow because diagnosis is slow. How correlated metrics, logs and traces, and alerts based on error budgets, took resolution time from four hours to thirty minutes.
Breaking up a monolith without a rewrite. The services are the easy part; data ownership and the arithmetic of failure decide whether it works.
The decisions beneath a design system: tokens named for intent, accessibility at the bottom of the stack, performance that is paid on every page, and adoption as the real product.