Our serverless superhero this week is Christian Bonzelet, Solutions Architect at DFL Digital Sports and AWS Community Builder. Christian not only designs highly-optimized high scale systems, but writes extensively about everything from serverless concepts to generative AI adoption on his blog. Thank you for everything you do for the community, Christian!
When building event-driven architectures, there are a few reasons why you’d need to capture all events flying through your system. You could need it for disaster recovery, end-to-end testing, or compliance reasons, to name a few. But what’s the best way to capture that traffic? Pubudu Jayawardana published an article last week telling us three ways to catch all events going through an EventBridge bus. He includes some cool options in there that make you think “why didn’t I think of that?”
Testing serverless applications is notoriously difficult - we talk about it all the time. Usually we talk about the reasons why it’s hard. But thanks to Matthew Wilson, we can now see some practical examples of how to do it. There are some interesting concepts in there that involve updating application code to account for tests, which might just be the key to effective testing in serverless production systems.
I wrote about a new automation I built into Ready, Set, Cloud last week. I found I was always forgetting to write posts on social media to announce when I had new content coming out. So I built a workflow that uses Amazon Bedrock to create, schedule, and post tweets for me. The post is a walkthrough of my thought process and the various components. If you’re interested in trying this yourself, I link to the source code in the article.
We’re seeing a surge in serverless capabilities (and popularity) in edge networks and browsers. So when I saw Wilmo Martinez’s post on adding HTTP security headers using Lambda@Edge I was intrigued. While the article doesn’t tell you how to add the headers, it does describe what the security headers are and what they do. It also shows in detail how Lambda@Edge is used with CloudFront as a CDN. It’s a helpful, educational read if you’re starting to venture into this space.
Lee Gilmore does great work. Every time he posts something new (which is like…all the time) I have to stop what I’m doing and read it. Last week he shared a forward-looking thought on AI as a platform for enterprises. He goes through in a ridiculous amount of detail how he sees AI panning out for global enterprises and I have to say - I’m convinced. The productivity boosts we’re going to see will be incredible with governance like what Lee has laid out.
Luciano Mammino and Eoin Shanaghy had Jeremy Daly on AWS bites to talk about a slew of serverless things ranging from serverless myths to the story of Ampt all the way to predictions about the future. It’s a super cool episode packed with info.
I had Michael Liendo on the Ready, Set, Cloud podcast to talk about the dirty truth of developer advocacy. We talk about everything from marketing funnels to how we decide on what content to create. This is a fun one!
Many of use Terraform for our IaC. But with the recent license change from HashiCorp there are lots of questions in the air about if we should still use it. Yan Cui had Anton Babenko on the Real-world Serverless podcast to talk about all of it. If you’re wondering what to do, tune in here.
Everyone in serverless knows DynamoDB, but how well do you really know it? Did you know there are some magic numbers that tell you if you’re optimally using it? Khawaja Shams shares a few incredibly cool numbers that tell you how to to find out if you’re paying too much for your current setup.
#DynamoDB is the most magical AWS Service, so @alexbdebrie & I decided to put on a magic show w/ @momentohq! #ddb_magic— Khawaja Shams (@ksshams) October 30, 2023
Our favorite DynamoDB magic numbers in🧵will help you become an expert table magician by learning key internals!
What's your favorite magic number? pic.twitter.com/5gC0ExwGUE
It was a light week for serverless releases. Not much to report on here besides being able to integrate with S3 Object Lambda while running Amazon Athena queries. Want to hide sensitive data at query time? No problem.
I gave a talk at the API World conference last week making the argument that all APIs should be serverless first. What’s your thought on the subject? There are numerous reasons to start new dev with a serverless architecture, can you think of any reasons not to? I’d love to hear them!