Node Interactive North America 2017

In software the act of logging is like breathing: you simply cannot avoid it. It is also (like one's own breathing patterns) quixotically nuanced. The expressions of this nuance can be seen in the number of ways it is used: in production server programs, in CLI tools, in web browsers, in native Mobile apps, just to name a few.

Winston is the most popular logging library for node. Released in early 2011, it is almost old as node itself. Over time there have been many other libraries that co-exist in the logging space: bunyan, bole, & pino. Yet the popularity of winston persists with over 6.5M monthly downloads and over 200 community contributed packages.

This talk will explore winston@3 and contrast it with other logging solutions that exist today. Not just from the perspective of performance, but from the perspective of flexibility across platforms.

Back pressure is a network protocol concept that is key to writing node services that perform well under load, and that don't accept more work than they are ready to do or store data without bounds if the data can't be acted on (yet). Scalability is a strength of Node.js, but requires careful construction of data flow in your application to break up CPU processing so multiple large requests can make progress simultaneously. Learn how streams support back pressure and how you can take advantage of it in your applications.

We all know that Node.js can scale, but how far and what's needed? Come learn how we scaled Node.js based AcmeAir to 30 billion transactions/day. The hardware, the software and the key challenges along the way that had to be overcome in order to get the 350,000 transactions per second! AcmeAir is one of the benchmarks used by the community to track Node.js performance, come learn more about this benchmark and how far it can scale.

In this talk I will cover the story of Node.js at Alibaba, a top Chinese Internet company consisting of many different subsidiaries, each with its own technical stack and business focus. I will talk about our effort to unify the Node.js development practices across the Alibaba group (frameworks, infrastructures, deployment, .etc), and how Node.js applications in Alibaba deal with the challenges of our Double 11 sales.

I will also talk about the open source Node.js projects from Alibaba (including egg.js and cnpm, two enterprise-facing projects born out of the unique environment of China), the public cnpm registry (funded by Alibaba Cloud, which serves millions of downloads from China per day), and the Chinese Node.js developer community.

GraphQL is seeing rapid adoption in the JavaScript community. In part, thanks to an excellent reference implementation written in Node.js (graphql-js). Yet there are still many challenges to operating a GraphQL API in production and few real world examples of how to tackle them. Newcomers are often on their own for concerns such as:

- Authorization and authentication
- Extending built in types
- Sharing field definitions
- Deciding between Union and Interface types
- Custom default resolvers
- Project structure and organization

In this talk I'll discuss why Bustle rewrote our entire backend in GraphQL and how we solved these challenges along the way. I will also be open sourcing code extracted from our production GraphQL backend.