Three petabytes a month

Node.js Testing Best Practices — A detailed guide to modern testing in Node from a group of developers who know all about it. It’s on GitHub but is essentially written like a free book covering over 50 battle-tested tips covering areas as diverse as the ‘Testing Diamond’, testing microservices, checking contracts, verifying OpenAPI correctness, and simulating flaky network conditions.

Goldberg, Salomon, and Gluskin

A Practical Guide to Llama Stack for Node Developers — Red Hat’s Node team has been digging deep into LLM coding in recent months, and now shares its discoveries in using Meta’s Llama Stack (not to be confused with their Llama LLM) – it’s a unified set of APIs for working with numerous parts of the modern LLM-powered stack including models, evals, RAG, tool calling, etc.

Michael Dawson

How the Node Team Makes Node.js Downloads Reliable — I’m sure most of us rely on just clicking and downloading Node.js (or having a package manager or CI process do it) and it Just Works™. Turns out there’s a lot going on behind the scenes to keep Node.js quick and easy to download and this tells the full story of an operation that serves up over 3 petabytes(!) of traffic a month.

flakey5

Could JavaScript Have Synchronous Await? — Our favorite JavaScript doctor explores the idea of a synchronous await to erase the divide between sync and async code, explains why it’s enticing, and digs into why performance hits and concurrency headaches make it impractical.

Dr. Axel Rauschmayer

Bare: A New Lightweight Runtime for Modular JS Apps — Imagine something like Node.js but really stripped back: bare, if you will. Like Node, it’s built on top of V8 and libuv, but Bare’s approach is to provide as little as possible (a module system, addon system, and thread support) and then rely upon userland modules that can evolve independently of Bare itself. It’s an interesting idea – more details here.

Holepunch