Together with Serverless, GraphQL rapidly changes the way we develop, maintain, and operate APIs to access data. Most of the current engineering techniques and well-known best practices will become obsolete with GraphQL and schema-first API designs. Using GraphQL requires a radical change in the way engineers think about designing APIs and how to structure data.
For the first time, it’s all about data access and consumers, context boundaries, and object relations. The API design is finally about what the consumer needs, not just what the server provides.
GraphQL is perfect for serverless architectures! Focus on your business logic instead of maintaining and operating infrastructure. Read more about Serverless Consulting …
With products like Prisma, you can replace your existing ORMs and access your data stored in MySQL, Postgres, or MongoDB with GraphQL. The auto-generated and type-safe client libraries simplify database access for your engineers; regardless of the preferred programming language.
Instead of operating your service, you can rely on services like the Apollo Platform. Based on the open source Apollo Client and Apollo Server, you get a fully managed platform for your GraphQL service; including a schema explorer, resolver metrics and insights, as well as support for improved tooling in your Continuous Integration and Continuous Deployment workflows.
The most radical solution to run a serverless GraphQL API is AWS AppSync. Together with AWS Lambda, you have a product that embraces the Functions as a Service mindest. Databases like DynamoDB and Serverless Aurora integrate so well into AppSync; you do not even need a Lambda function with custom business logic to query your data anymore.
Together with my colleagues at superluminar, I offer workshops, training, and architecture reviews for all things serverless and GraphQL. Regardless if you want or need to prepare your application for the future, you should talk to the industry experts in Hamburg for running GraphQL APIs on Amazon Web Services.
Avoid running into the same pitfalls others already ran into. Do not waste time with solving problems somebody else already solved and design your next data API for the least amount of maintenance and operational tasks. Just say hi …