November 2023 Summaries

This guide teaches how to create a Serverless API using AWS Lambda, CDK (with TypeScript), and Neon. The API is a REST API with a single HTTP endpoint that returns a list of products from a Neon database. The prerequisites include basic knowledge of JavaScript/TypeScript as well as Node.js, an AWS account, and experience working with AWS. The guide covers the installation of CDK CLI, setting up a new CDK project, defining Lambda function, app stack, creating a Neon Postgres database, storing its connection string in AWS Secrets Manager, deploying the project, and cleaning up afterwards.

This guide provides a comprehensive tutorial on building a serverless API using AWS Lambda, AWS Cloud Development Kit (CDK) with TypeScript, and Neon, a serverless Postgres database. It outlines the process of creating a REST API with a single HTTP endpoint that retrieves a list of products from a Neon database, emphasizing the need for basic knowledge in JavaScript/TypeScript, Node.js, and AWS. The guide details the use of various technologies, including AWS Lambda for scalable, serverless code execution, and AWS CDK for defining cloud infrastructure as code. It explains how to set up a development environment, create and deploy a CDK project, define a Lambda function, and securely manage database credentials with AWS Secrets Manager. Furthermore, it covers the deployment of the API, retrieval of database connection details, and provides instructions for cleaning up resources to prevent incurring unexpected AWS charges. The guide concludes with links to additional resources and articles for further exploration.

Neon enables developers to create isolated preview environments using its database branching feature, which allows teams to create a copy of their project's Postgres cluster with unique branches that are completely isolated from each other. Kubernetes and Argo CD can be used as hosting and continuous delivery platforms to automate the deployment of these preview environments. A GitHub Actions workflow is also utilized to manage the lifecycle of the container image and Neon resources, including creating a new database branch and compute endpoint for each pull request. This setup enables developers to quickly test changes in an isolated production-like environment without impacting their main production system.

The blog post explores the creation of preview environments using Neon, Kubernetes, Argo CD, and GitHub Actions, enabling developers to test code changes in isolated, production-like settings automatically generated in response to pull requests. It highlights Neon's database branching feature, which allows the creation of separate, isolated branches of a Postgres cluster for efficient, cost-effective testing without impacting the main production system. The setup involves using Kubernetes for scalable compute clusters and Argo CD for automated application deployment, with a GitHub Actions workflow managing the lifecycle of container images and Neon resources. This approach allows for continuous integration and delivery, facilitating rapid feedback and testing cycles, and provides a practical example of integrating serverless architectures with modern CI/CD pipelines.

A serverless database refers to a cloud-based service where users can store and manage data without worrying about managing servers, scaling resources, or handling administrative tasks. This approach is designed to eliminate the operational burden of traditional databases, allowing developers to focus on their applications' functionality. Serverless databases are highly scalable and resource-efficient, as they automatically adjust to changing workloads and only charge for the actual usage. Neon, a serverless Postgres platform, aims to make databases just a URL, simplifying management and reducing costs for both individual developers and large organizations.

Serverless databases, as highlighted in the article, represent a shift from traditional serverfull architectures where developers had to manually manage resources like CPU and memory, to a more automated system where these concerns are abstracted away. This transition is akin to the move from physical servers to cloud computing, offering near-instantaneous scalability and addressing the common issue of over-provisioning resources, which often led to unnecessary expenses. Neon has developed a serverless database platform for Postgres, emphasizing the elimination of operational burdens such as maintenance, scaling, and resource management, allowing developers to focus on service delivery. This model proves advantageous for managing large-scale deployments, particularly for individual developers and organizations, by reducing costs and simplifying database management. While serverless functions have their limitations, serverless databases are positioned as a long-term solution offering resource efficiency and operational simplicity, with Neon positioning itself as a key player in this field.

NPM is a package manager for JavaScript that assumes it's running on Node.js, causing issues when used outside of Node.js. To resolve this, developers use techniques such as third-party shims to replace Node-specific features with equivalent libraries, bundling code with esbuild to combine dependencies into a single bundle, and creating stub packages or re-implementing critical functionality. These methods enable NPM packages to be run in environments beyond Node.js, including Vercel Edge Functions, web browsers, and serverless platforms.

George MacKerron illustrates techniques to run NPM packages outside of Node.js environments, such as Vercel Edge Functions and web browsers, using esbuild to create runnable bundles. He highlights the challenges posed by Node-specific features in NPM packages, which can hinder their use on non-Node platforms. MacKerron shares strategies like using third-party shim libraries for global objects, creating stub packages for less critical Node libraries, and re-implementing Node features for essential functionalities. He also discusses package masquerading to substitute dependencies, demonstrating these methods in the development of the @neondatabase/serverless driver, adapted from the node-postgres pg package, to make Postgres usable on serverless platforms.