Blog
Notes on how Fairyfly works, the choices behind it, and what changes between releases.
Benchmark · Telemetry
Serverless MicroVM Benchmark: Cold Starts vs. 5M Requests at 500 Concurrency
An honest telemetry breakdown of isolated cold starts (~10ms inside-VM boot) and warm sustained throughput (165k RPS across 5M requests) across Fairyfly, Bun, Deno, and Node on 2 vCPU / 1 GB RAM microVMs. Read the benchmark →v0.2.0-beta · Design deep dive
What is Fairyfly? How it works and why it works that way
A tour of the serverless-first design, the four layers, the single event loop, the HTTP hot path, and the deliberate limits that keep Fairyfly small, fast, and auditable. Read the post →More posts will land here as releases ship.