from Hacker News

After 10 years of running a web development agency and delivering over 100 projects, we tried many different approaches to building apps and websites. When we looked at the patterns for most of the projects we delivered, it was always a combination of custom business logic and a way to manage content, like pages, news articles, and similar. Looking at the options available on the market, we either had frameworks for building the custom logic or ready-made CMS solutions for managing the content. There was no combination for when we wanted to do both. And this is one of the biggest pains we had. We would either force custom logic inside a CMS and break things or make them hard to maintain, or use a framework and take much longer to deliver a project since we'd end up building a custom CMS and making the whole thing more expensive to the client. On top of that, we were just tired of constantly spinning up servers, managing container images, worrying about uptime, network, and security issues, and paying for resources we were not utilizing 100%.

Looking for options, we discovered serverless. The premise of not having infrastructure to manage sounded really intriguing. Having fault-tolerant resources that scale automatically when you need them with consumption-based pricing that cost up to 80% less than virtual machines sounded like the ideal solution...until we tried to build something with it. It was almost impossible. All the existing frameworks and CMS options are designed for a "server environment", and couldn't be used to build solutions in a serverless environment. The only tutorials available at that time covered how to resize an image with a Lambda function. Besides that, serverless requires a cognitive change of how you approach code and infrastructure.

In all those problems we saw an opportunity. Over a course of a year, we built a framework that allows anyone to quickly build serverless applications without battling all the challenges that come with it. Things like rendering and caching pages, optimizing the cold-start times, debugging function calls, managing connections between functions and the database, CI/CD setup, and many more. The framework comes with a GraphQL API, Admin UI, ACL, CLI for deployment and scaffolding, and more.

Because our passion is also tied to content management systems, we decided to eat our own dog food and build a serverless CMS using our own framework. Webiny Serverless CMS uses Lambda functions, API Gateway, DynamoDB, Elasticsearch (the only non-serverless component at the moment) and S3. It scales automatically together with the demand, requires minimal maintenance, and costs a fraction when compared to solutions running on virtual machines or containers. It includes a GraphQL API, asset management, and a no-code builder for static pages and forms. It takes 2 commands to install and configure the whole thing. Today it runs only on AWS, but we plan to introduce the support for other clouds in the future. With our CMS, we hope people will be more confident in the abilities serverless brings to the market, especially when building full-stack solutions. To put some numbers behind our product, we made a benchmark to demonstrate the performance and cost of a Webiny full-stack serverless application[1].

Both the framework and the CMS are free and open-source[2] under the MIT license. We do have a paid enterprise offering for those that require support and additional features [3].

Give Webiny a spin: https://www.webiny.com/. We would love to know what you think!

Resources:

[1] Benchmark - https://docs.webiny.com/docs/webiny-overview/performance-ben...

[2] Github - https://github.com/webiny/webiny-js

[3] Pricing - https://www.webiny.com/pricing