Our serverless superhero this week is Michael Bahr! Michael is a phenomenal serverless developer with an impressive track record. He posts great content on his blog and even publishes a serverless newsletter. He’s available for hire! If you’re looking for a rockstar, give him a shout. Thanks for everything you do, Michael!
I’m a big fan of event-driven architectures. The patterns that emerge as a result of loose coupling your resources are innovative but can sometimes be tricky to really understand. Elias Brange wrote an article talking about the transactional outbox pattern for serverless domain events last week. I love this one, it has clean domain separation, clearly states a naive vs production-ready approach, and discusses the all-important error handling component to the pattern. Well done, Elias!
Ran Isenberg posted a great article on how to effortlessly monitor serverless applications with CloudWatch. He covers a lot of ground in this article, but I got the most value out of his guidelines on what to focus on for building dashboards and how he makes a distinction between high-level and low-level dashboards.
I love a good migration story. I often read the successes of a cloud migration going from on-prem or containers to serverless. But I rarely see migration stories in serverless from one programming language to another. Min Jie Fan shared the story of how his company migrated from Python to Typescript in their Lambda functions. He gives his point of view on the tradeoffs of the two languages, weighs in on cold starts, and shows the approach of how they refactored one function at a time.
Lee Gilmore is back and this time turning his production-readiness eye toward Amazon Bedrock and using a CLI to generate clean code. This is a fascinating approach to leveraging AI to do your dirty work. Just give it some definitions around what you want to build and you have strongly defined schemas and types. Very cool and lots of potential approaching development with AI this way.
I’ve been following a great series from Daniele Frasca on serverless latency as you scale. His latest article describes the challenges of scaling using Redis as a read-aside cache. He describes the difficulty, implementation, and design decisions in great detail and my key takeaway is - this does not seem like a fun thing to do.
I wrote an article talking about four real-time notification mechanisms and how they are different from each other. My goal was to shed some light on the options you have available when communicating between parts of a distributed system. It’s way more than just WebSockets.
Congratulations to Eoin Shanaghy and Luciano Mammino on their 100th episode of AWS Bites! This episode focuses on Ampt, discussing everything you’d ever need to know about this shining star in the field.
I had Cliff Crosland on the Ready, Set, Cloud podcast to talk about how Scanner built an ultra-fast serverless data lake. We talk about what data lakes are, but more importantly we talk about Cliff’s serverless design that enables petabyte scale queries in seconds. This is one of the coolest serverless designs I’ve seen in a while.
Amazon CloudWatch announced out-of-the-box best practice alarms for 19 AWS services. This was a big gap in the observability space. I love it that we can use this to tell us what we need to know about our applications.
Baselime has expanded its capabilities by adding support for logs on Vercel and Cloudflare, along with distributed tracing via OpenTelemetry for Next.js. These enhancements offer more comprehensive observability for your applications beyond the AWS serverless ecosystem.
Anyone who uses DynamoDB has like heard of single table design. It’s something that is advertised to the masses as the way to build your NoSQL data models. But should it be? Pete Naylor wrote an incredible article describing why single table design is not necessarily always your best bet and explains in detail what could go wrong when you try to use it where you shouldn’t. It’s an eye-opener for sure.
This isn’t serverless related, but a post caught my eye last week on what I believe to be (mostly) true in the tech industry. It’s a simple image that might understate the importance of a college degree, but definitely highlights the fact that you can get into tech without one. Practice, practice, practice!
As a programmer - write code everyday! pic.twitter.com/MwLInWXEfc— gbae the app developer (@daboigbae) October 19, 2023
Serverless Days Cape Town was last week, bringing in some of the best serverless minds around the world. Stay tuned while we wait to see the talks come online.