Serverless Superhero

Our serverless superhero this week is Vadym Kazulkin, head of development at ip.labs GmbH and AWS Community Builder. Vadym has been involved with the serverless community for years and regularly shares and creates content in the space. He was recently on the Ready, Set, Cloud podcast and also gave a talk at the AWS Community Day DACH last week. Thank you for everything you do, Vadym!

Tutorials

In a continuation of his series around user-based auth, Arpad Toth is back with a killer article on access control on API Gateway with ID tokens. I continue to learn so much about how to pass along permissions via Cognito to downstream services. This is a must-read for anyone looking to implement this type of authentication!

I’m not usually a big fan of step-by-step code tutorials but this one from Elias Brange is on such a good topic and explains the concept so well, I just had to share it. He talks about how to do canary deployments of Lambda functions with SST. I’ve tried to do canary deployments before with CodeBuild and let me tell you, it’s not an easy setup. Elias explains how to do it simply and why it’s important to build into your production applications.

I really like the thought behind Daniele Frasca’s article last week, is serverless scaling achievable without caching. He talks about issues with serverless applications when they reach a certain amount of scale, shows how he gets throttled by various AWS services on his way up, and demonstrates how he mitigates some of the problems. I’m very much looking forward to the last part of this series!

Benjamen Pyle wrote about how to process DynamoDB streams with EventBridge pipes with multiple items last week. If you follow Single Table Design or even just keep related entities in the same DynamoDB tables, you have to account for that in your streams. Benjamen walks you through how to do it and even makes a case that pipes could be replacing some Step Function usage for him!

New Releases

AWS AppSync WebSocket subscriptions now support nested filtering. This feature comes in with no data model changes on your part, so you can start getting better filtering today.

The API Gateway console has gotten a facelift! A much-needed UX improvement was released last week giving a significantly friendlier experience to the console.

SNS FIFO topics can now deliver messages to SQS standard queues. This is a great update that alleviates constraints put on many of us with the design of some of our systems in the past. Of course, this doesn’t mean guaranteed ordering in the SQS queue but it’s nice to be able to fan out and work multiple jobs off of SNS messages where some require ordering and others do not.

Spotlight

The internet has been buzzing about Bun, the new(ish) alleged competition to Node.js. Will it work in serverless environments? Will it be better than Node? Mitchell Kossoris did the early test for us, putting both frameworks through a benchmark on Lambda for IO- and CPU-bound tasks. As we know, we’re early for this framework, but I was surprised to see the results!

Tip of the Week

I’m a big fan of Luciano Mammino’s podcast and I particularly liked this episode debating if you should mount an S3 bucket as a file system.

As promised, we are back with a new episode of #AWS Bites #podcast!

Today we talk about mounting S3 buckets directly in your filesystem using AWS Mountpoint, an interesting open-source project written in #Rust.

Is it a good idea though? 🤨https://t.co/JBU1kRh2Xj

— Luciano Mammino 𝕏 Node.js Design Patterns (@loige) September 15, 2023

Last Words

There’s so much talk about TypeScript vs JavaScript, Bun vs Node, etc. What do you think, does it really matter? Seems like people are saying as long as you get the job done efficiently and well, then the means aren’t important. Do you agree?

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