I Trust Boring Infrastructure More Than Clever Infrastructure
Clever infrastructure looks impressive in diagrams. Boring infrastructure is usually easier to operate, easier to debug, and much easier to keep alive once real users depend on it.
wade womersley – york based software engineer
Engineering notes from the sharp end
wade womersley – york based software engineer I write the way I work: direct, useful, and more interested in what holds up in production than what sounds clever on a slide.
193 published posts
Latest update
Latest post
Clever infrastructure looks impressive in diagrams. Boring infrastructure is usually easier to operate, easier to debug, and much easier to keep alive once real users depend on it.
Archive
Environment management usually looks simple at first. Then defaults drift, secrets spread out, local setups stop matching production, and the whole thing becomes a source of avoidable mistakes.
AWS diagrams love to look simple. The problem is that the operational reality behind them is usually doing a lot more work than the picture admits.
Serverless has real tradeoffs, but for small teams I still think it usually wins. The operational overhead stays low, the first version ships faster, and the mistakes are easier to afford early on.
I still like serverless, but the tradeoff is obvious when something breaks at 2 a.m. The architecture is easy to ship and harder to reason about when you need logs, context, and a fast path to the real failure.
Clients do not really need a consultant who just says clever technical things. They need someone who can reduce risk, create clarity, and help them make better decisions.
A legacy system is not dangerous just because it is old. It becomes dangerous when nobody can change it safely, nobody can see what it is doing, and the business depends on it anyway.
AI can generate code faster, but that only makes direction, judgment, and accountability more important. Senior engineers still matter because the expensive mistakes are rarely typing mistakes.
Most rewrite discussions are really a sign that the team is frustrated, not that a full rebuild is the smartest technical decision. In most cases, incremental change is the better call.
AI tooling is not going away. What a lot of people are calling a bubble looks more like denial in the face of tools that are already becoming normal developer infrastructure.