Technical Insights

Most of it centers on the Rails 8 production stack: running Solid Queue, Solid Cache, and Solid Cable in place of Redis and Sidekiq, deploying with Kamal, and building UIs with Hotwire. There's a fair bit on databases too, since that's where the interesting failures tend to surface: PostgreSQL tuning, connection pooling, and reaching for TimescaleDB once the metrics tables outgrow an ordinary query. Underneath both sits the architecture question, like when a monolith still beats splitting things apart, where a service object actually earns its place, and how to model a domain without burying it in abstractions you regret later.

The business side runs through all of it, because cost, speed, and risk are what a client is really weighing when they pick one approach over another. Everything here comes out of production work, so you get the real numbers and a straight account of where each choice breaks down.