From Experiment to Backbone: Adopting Rust in Production (opens in new tab)

We originally introduced Rust (back in 2018) as a small experiment alongside existing systems, mostly to validate safety and performance assumptions under real production load.

Over time, the reduction in memory-related incidents and clearer failure modes led us to expand its use into increasingly critical paths. This post focuses less on “Rust is great” and more on the tradeoffs, mistakes, and organizational changes required to make that transition work in practice.

Happy to answer questions about what did not work, where Rust was a poor fit, or how we handled interop with existing systems.