Why I Needed Rust, Finally
Infrastructure automation I can sleep on — Rustikon 2026, Warsaw
Why I Needed Rust, Finally
I gave this talk at Rustikon 2026 (Warsaw, March 19, 2026): Why I Needed Rust, Finally — Infrastructure Automation I Can Sleep On. It’s the story of a long resistance and of what finally broke it.
The resistance
For years I automated infrastructure with the usual tooling: scripts, templates, layers of YAML that validate late — once they are already running against a real environment. It worked until it didn’t, and the failure always showed up at the worst possible moment: at deploy time, never before.
What broke it
Rust didn’t win me over on performance. It won me over because it moves failure to the left: from deploy to the compiler. A type that doesn’t line up never reaches production — it doesn’t even run. That guarantee, applied to the provisioning layer, is what turns “I hope this deploys” into “this cannot deploy wrong in that way.”
Infrastructure you can sleep on
The title isn’t a slogan. When configuration is the source of truth and it’s typed, and the system refuses to start if something doesn’t fit, being on call stops being an act of faith. That was the point in Warsaw: not that Rust is fast, but that it lets you sleep on it.