In this day and age, dealing with REST endpoints is a requirement — but don’t let that stop you from using GraphQL!
Jon Wong, a Senior Software Engineer at Coursera, teaches how the Coursera engineering team takes advantage of REST to build out GraphQL servers and explains some of the strategies learned along the way. In addition, he presents about some of the tools, libraries, and workflows that make layering GraphQL on top of REST APIs a breeze.
. . .
Deploying a Serverless GraphQL API
More and more, we find that GraphQL APIs are being developed by product developers that don’t want to worry about operating backend services.
The serverless function model of deployment has a lot of benefits for APIs, but isn’t as well suited when you need some stateful features such as caching or performance metrics aggregation.
In this talk, Sashko Stubailo, the Open Source lead at Apollo, presents a solution that combines the best of both worlds: A stateless GraphQL API running in AWS Lambda for ease of scaling with Apollo Engine in front to handle stateful concerns.
. . .
Client-Side Caching with Apollo Link
John Haytko, a Software Developer at Xero, explores the current state of client-side caching with Apollo and how Apollo Link can make managing your client-side cache easier and more scaleable.
. . .
SDL First Products
Christopher Biscardi, a Product Engineer at Honeycomb.io, describes how the GraphQL SDL can be used to decouple client and server implementations and why that matters.
He covers a set of tools that are “Schema First”, including Prisma and Apollo, as well as dives into how the SDL can be used to communicate across teams.
. . .
End-to-end Type Systems: More Features and Fewer Bugs
Stephen Wan, an Infrastructure Engineer at Samsara, describes the necessity of static typing everywhere at Samsara: Protobufs, Go, GraphQL, and TypeScript.
In this talk he discusses the uses of Thunder, Samsara’s Go GraphQL server, apollo-codegen, and other tools to automatically glue these pieces together and make development easier with close to zero human intervention.