Serverless Superhero

Our serverless superhero this week is Omkar Kadam, lead DevOps engineer at Cactus Communications and AWS Community Builder. Omkar recently published a book called The DevOps Story where he shares his knowledge on the evolution and principles of DevOps. He is also an active community member, regularly engaging with individuals online offering his expertise on everything serverless and DevOps. Congrats on the book and thank you for everything, Omkar!

Tutorials

How many of you have created an API without any form of auth in front of it? Yeah, me too. Turns out, it’s not that difficult to secure your APIs - even the ones for your dev experiments - with Cognito. Andres Moreno published an article last week describing how he secures his APIs with a single Cognito user pool. This is great for reusability across projects and limiting the number of resources you have in the cloud. Andres explains what each piece is in his setup and shows how to get started using SAM.

Ben Pyle wrote a super useful article last week describing how to deploy WASM on S3. This is part two of a short series on serverless WASM with Rust, and the wrap-up to get you going with lightning-fast static sites. Ben walks us through how to set up an S3 bucket and CloudFront distribution to serve the data and also shows us how to compile Rust code down to WASM using Trunk. I’ll be honest, a lot of this feels over my head, but the way Ben presents it makes me encouraged to try it out.

Interesting Content

I had Danielle Heberling, James Eastham, Andres Moreno, and Ben Pyle on the Ready, Set, Cloud podcast last week to talk about the Believe in Serverless community. We talk about how the community has been growing, what our favorite moments have been, and where we see it going the rest of the year. I’m really excited about this community and hope you all check it out. It’s full of wonderful people who sincerely want to see each other succeed.

Data streams can be intimidating if you’ve never built with them before. So can the Rust programming language. But somehow James Eastham makes both of them seem… easy. In his video last week on processing Kinesis streams with Lambda and Rust, he provides an engaging explanation of how to make blazing-fast stream processors, how to handle errors, and how to simulate data ingress in your projects. I love this video so much, it’s so well done and covers an extremely valuable topic in software development.

Jeremy Daly did a talk for the Believe in Serverless community last week covering self-provisioning serverless with Ampt. He walks through the extremely cool stuff that Ampt does for you to make building quick and easy. Then, I might have thrown him a curveball when I asked him to build a meme generator using the integrated AI capabilities of Ampt. I won’t spoil how it went, you’ll have to watch to find out. Since I know he reads this newsletter - thank you for spending the time to teach us about Ampt, Jeremy. It was great!

Spotlight

The difficulty of managing dependencies in a large application is underrated. I remember struggling with how to scale IaC across development teams years ago when I was starting a platform team at a previous job. Honestly, I never really found a great answer. That said, Lee Gilmore might have found one. In his article last week, he shares two approaches to managing AWS CDK stack dependencies. He describes the differences between using dependencies as references and importing dependencies across stacks and gives you meaningful code and explanations of each. This is a must-read before your application and practices scale up - it will save you lots of heartache!

Tip of the Week

We often think of the different types of software tests as a pyramid. A large amount of unit tests make up the base of the pyramid, integration tests make up the middle, and end-to-end tests make up the top, with each layer of the pyramid having fewer and fewer tests. But Yan Cui reminded us last week it might not be that way with serverless. It might be more of a honeycomb shape for testing.

Serverless apps have a different risk profile to serverful apps. Much of our app is outside of our code as we delegate many responsibilities to managed services.

That's why the "test honeycomb" is a much better fit for us, check out this 📽️ to learn more: https://t.co/bk8kNxw5XO pic.twitter.com/ISPOYnyjaW

— Yan Cui (@theburningmonk) May 11, 2024

New Releases

Amplify Gen 2 is generally available. I tried this out the other day and can say with confidence - this is an extremely easy way to build full-stack applications. I love how easy the team has made it to build small to medium-sized apps.

AWS announced Bedrock Studio in preview last week. This is a visual workspace intended to help you build GenAI applications. I haven’t tried this yet, but it looks promising!

Amazon Polly received three new synthetic voices to debut a new generative engine. These voices are allegedly highly expressive and are supposed to sound more human than ever.

Cognito announced tiered pricing for machine-to-machine use cases. Base pricing did not change, so it just got less expensive for high-usage workloads.

Last Words

Don’t miss the upcoming free virtual conference, Moar Serverless!! 2024 on May 23 from 11-5pm EDT. This conference has a wonderful lineup of speakers and is sure to spread some amazing concepts in serverless.

If you’d like to make a recommendation for the serverless superhero or for an article you found especially useful, send me a message on Twitter, LinkedIn, or email.

Happy Coding!

Allen