Serverless Superhero
Our serverless superhero this week is Steve Sanders, CTO at DonorSpring. Steve is an active member of the Believe in Serverless community, regularly offering his insights and experiences on everything from EDA to front-end tips. He also works on really neat open source side projects like episode ninja that are great resources to learn from. Thanks for all you do for the community, Steve!
Interesting Content
When I think of event-driven architectures, I often think of James Eastham. He is one of the few content creators out there today that explains EDA in a way that is approachable to beginners and advanced practitioners alike. His recent talk from AWS ComSum about observability in an asynchronous world was published on YouTube last week and one word comes to mind: wow. It’s a fantastic talk that needs to be on everybody’s watch list no matter how much you’ve worked with events. Fantastic job, James!
We’ve all heard the infamous “vendor lock-in” argument when it comes to serverless. Seems like nobody wants to make a commitment. For whatever reason, that logic doesn’t seem to apply for many people when thinking about savings plans and long-term commitments to lower cloud spend. You sign up for 1 to 3 years and get a cheaper price as a result. To me, that’s way scarier than opting to choose Lambda vs a container. With Archera Insured Commitments, you lock in discounts without the risk — if your needs change, they cover the extra cost, so you aren’t paying for resources you don’t use. It’s like having the best of both worlds: cost savings without getting stuck in rigid, long-term deals. Learn more. Sponsored
Khawaja Shams dropped some gold on us last week. He wrote a blog post about how S3 is amazing, but not as a live media origin. If you don’t know what a live media origin is - don’t worry, Khawaja explains what that is beautifully. He also has that really cool math you didn’t know you didn’t know and when you see it in front of you things just… make sense. I love this article so much. The ethos of it being “use the right tool for the job.”
Strong opinions are always fun. Mostly. Ok sometimes they can be a total disaster. But that was not the case with Pete Naylor last week when he talked about the state of databases in 2024. For those of you that don’t know Pete, he’s the one you want to listen to for topics like this. Pete was on the DynamoDB team for years during a pivotal time for the service. He now works at EDB and is neck deep in Postgres. The talk itself was great, but my favorite part were the hard hitting questions at the end from the live viewers.
Alongside strong opinions are of course… unpopular opinions. Which is exactly what Eyal Estrin called his article about moving back to on-prem. The article talks about staying in the public cloud and how he views it as the future for the tech industry. There’s a lot of what we’ve heard before for arguments for the cloud, but Eyal has a couple of points in there I hadn’t seen before and thought were really insightful.
Spotlight
I’m not exactly sure how he does it, but AJ Stuyvenberg always seems to find super important issues with how Lambda works. Last week he shared a blog post on something he calls the Lambda doom loop which has started cropping up as a result of a recent behavior change in Lambda. Apparently subsequent initializations of a Lambda function now count toward your overall timeout, which can lead to a devastating effect if not handled appropriately. AJ explains the issue and offers four ways to get around it in his article.
Tip of the Week
The Believe in Serverless party is coming back to re:Invent!! Momento just teasted the party with a fun little highlight reel of last year’s event. Can’t wait to see you all there!
Momento is bringing back the #BelieveInServerless party at #AWSreInvent. Be part of an unforgettable night celebrating the @believeinsls community, AWS Heroes, 10 years of AWS Lambda, and YOU! Join our waitlist to save your spot! 💚 https://t.co/aeT1IxDx3P pic.twitter.com/PPtiLChp1m
— Momento (@momentohq) October 11, 2024
New Releases
OpenSearch Serverless got a bunch of updates last week, including a flat object data type, faster indexing and search performance, support for geospatial features, and surprisingly a lot more.
Valkey got some serious love last week with ElastiCache and MemoryDB support. Both services are seeing some serious price drops as a result. It’s a great step forward for caching, but you know where my heart lies when it comes to actual serverless caching 💙
Last Words
Halloween is coming up! Right now I’m talking to people from the tech community to get their scary production horror stories. If you’re interested in telling me about what went horribly, horribly wrong, send me a DM or reach out on Discord. I’d love to learn about it and help you share the story with others!
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
This Week's Sponsor
