Serverless Superhero

Our serverless superhero this week is Seth Orell, principal engineer at Parsable and AWS Community Builder. Seth is a tremendous serverless advocate, frequent contributor to the community, and great blogger. I love seeing his insights from years of deep technical experience shine through in the community. Thank you for everything, Seth!

Tutorials

You can never talk about security enough. Pubudu Jayawardana shared a blog post last week talking about encryption options with SQS. But his article is different from SQS encryption posts we’ve seen in the past. He talks about how you can manage the different types of encryption across AWS accounts. This is a heavily used use case but something we barely see any content about. Pubudu explains the permissions necessary with examples of how to protect your data using different types of encrpytion. Awesome post!

Darryl Ruggles showed us how to create an audit bot in your AWS account that monitors for specific actions. This bot uses CloudTrail to watch for certain activity in your account, like deleting an S3 bucket or logging in as the root account user, and then sends you an email and slack message saying something fishy is going on. It’s a surprisingly simple build with a huge safety impact on an organization. Very clever and practical!

More and more discussions around Cloudflare have been popping up lately. I’ve been seeing huge promise around the speed and functionality of Workers compared to Lambda. James Eastham published a video last week on how to build your first Cloudflare Worker with Rust. The video is much more than watching him code, it explains how Cloudflare approaches FaaS, how to setup your dev environment, how to deploy, performance testing, and yes - coding in Rust. This is a complete video that will leave you ready to go build one yourrself.

Interesting Content

Last week, Andres Moreno and I wrapped up our “can you actually build something real with AI assistants” series by trying out Amazon Q. Our goal was to build a GPS data ingestion app that runs analytical queries after every 10 data points. We directly compared the results with the code generated from Claude artifacts the week before and were… unpleasantly surprised at the quality. Between losing file history after each code iteration to having our prompts ignored, I think Q has a long way to go before it’s ready for prime time - at least it’s agent functionality. The stream is worth a watch, it was indicative of its current capabilities and pretty funny if I’m being honest 😊

I discovered the Lambda Live Debugger last week created by none other than our very own Marko. This project does exactly what you’d think, but in a super cool way. This will allow you to debug Lambda functions in the cloud directly on your local machine. It intercepts and routes traffic in the deployed function to your machine via AWS IoT and returns the response back to Lambda. It’s a really neat project that can help you with troubleshooting issues when you need it. It’s been a while since I’ve line by line debugged, but I can see it being useful in a production scenario where resolution time is critical. Best of all, it’s free to use! Just be wary of the increased Lambda costs since your execution time will be skyrocketing (also watch out for timeouts as well).

Spotlight

There’s a lot of kerfuffle around serverless. Always has been, but as of late the advocacy against it seems to be getting stronger. The arguments haven’t changed, but the narrative around it isn’t being challenged like we’ve seen in the past. That’s why talks like the one Danielle Heberling gave last week are so important. In a session called Rethinking Serverless, Danielle talks about everything you’ve ever heard about serverless - cold starts, vendor lock-in, trade-offs, complexity, limitations, and more. It makes you feel reinvigorated and ready to go build something new. My favorite part of her video: do something > complain.

Tip of the Week

When I first started with DynamoDB, I found it surprisingly difficult to implement pagination. Yan Cui shared an oldie but a goodie last week on exactly that topic.

Student: I need to implement pagination with DynamoDB, help!

Me: Don't worry, I've got you covered! This article covers both unidirectional and bidirectional pagination 😎https://t.co/VfrcWGRLHK pic.twitter.com/P9kvoaIAAH

— Yan Cui (@theburningmonk) September 7, 2024

New Releases

Congratulations to the new AWS heroes!! I’m so proud and happy you are getting the recognition you deserve for your tireless, selfless efforts. Special shoutout to new serverless heroes Lee Gilmore and Jimmy Dahlqvist!!!

DynamoDB just got attribute based access control (ABAC), allowing you to restrict access to records based on a tag on a user, role, or AWS resource.

Amazon Bedrock now supports Stability AI’s best models - Stable Image Ultra, Stable Diffusion 3 Large, and Stable Image Core. From what I hear, these are some of the best models for developing photorealistic images.

Bedrock agents now have the option to use Claude 3.5 Sonnet. From my personal experience, this is going to be a huge boost in quality.

Last Words

Over the past few weeks I’ve been trying out serverless or managed services that are NOT from the big 3 (AWS, GCP, or Azure) and I am loving it! Usually these services are actually built on top of the big 3, giving us the infrastructure reliability we love, but we’re getting the deep technical expertise from dedicated service teams. Products like Momento, Neon, and now Typesense are all rocking the tech landscape by providing niche services that are solving problems better than we’ve ever seen before. My willingness to try something that is not AWS has skyrocketed because I’m getting huge boosts in developer experience and dev time. Are you up for trying new services? Why or why not? Let me know!

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