Serverless Superhero
Our serverless superhero this week is Gary Sassano, senior DevOps engineer and AWS Community Builder. Gary is a fantastic open source contributor and is heavily involved in the community offering his insights and opinions to others. Thank you for all your contributions and sharing your expertise, Gary!
Interesting Content
When it comes to deployments with serverless architectures, many of us take the simplicity of Lambda for granted. We (I assume we, definitely me) use the default deployment settings in our IaC, make changes, and push. But what if you pushed a bug and you have hundreds of thousands of users? Lee Gilmore shared an article last week about progressive Lambda deployments with rollbacks which address this problem. Instead of “big bang” releases, you shift traffic little by little and watch for errors. Lee presents this solution in a simple, pragmatic way and of course, walks you through in detail with a functional code sample.
Do you like the thought of steeply discounted savings plans but aren’t sure if you’ll use everything you’re paying for? Fear not, with Archera’s Insured Commitments, you’re completely covered. Get a real, handwritten, paper check for the difference in your usage plan if you underutilize the resources you pay for. That’s a fancy way of saying “get commitment prices without the commitment.” Sponsored
Remember the last time you accidentally deleted something then instantly thought to yourself, “oh no, I shouldn’t have done that”? I hope I’m not the only person that happens to. Anyway, I really like the article from Halil Ural last week where he talks about implementing soft delete with DynamoDB. The concept is not new nor is the implementation, but it’s a great reminder that a delete doesn’t have to be permanent. By using attributes and filters on your database items, you can temporarily persist “deleted data” until a TTL expires. Give it a read and see if you want to bring that pattern into your apps.
I absolutely love novel approaches to explain things and educate others. Last week, David Behroozi shared a super fun and engaging way to learn about his product, Speedrun. On the Believe in Serverless stream, David and Danielle Heberling played an escape room challenge with Speedrun showing it’s incredible capabilities and how it can really boost your productivity. I love the idea and love the implementation. Great work, David!
I stumbled upon a really interesting video last week from NeetCodeIO that went into detail about how microservices went horribly wrong at DoorDash. DoorDash is a big, successful company that everybody has used at least once, and they’ve been able to get where they are with a monolithic architecture. When they moved to microservices, or in their case a “distributed monolith”, they were flooded with gotchas that not too many people talk about. Well, this video talks about them and explains them incredibly well. While this isn’t directly serverless, it is about a concept many of us practice.
Spotlight
If you’re actively involved in the tech community even in the slightest, you’ve probably recognized the disenchantment with generative AI over the past several months. Advancements in tools and services we aren’t using in production are being announced weekly and the big cloud vendors seemingly are pouring the majority of their efforts into outpacing one another. The latest evidence of this was brought to light by Mike Graff who wrote a speculative by-the-numbers article that asked the question “has generative AI ruined the re:Invent experience?”. With AWS releasing the session schedule last week, Mike made quick work of figuring out what the focus is on this year. I’ll give you one guess on what that is.
Tip of the Week
When I first read this tweet from Vercel, I couldn’t stop rolling my eyes. “Serverles servers”?! But then I actually read the article and am impressed at the engineering behind this cringy phrase. Vercel seems to have tackled a a huge “problem” with Node.js Lambda functions. I’m curious to hear what people think about this one.
Serverless servers: Node.js with in-function concurrency
— Vercel (@vercel) October 3, 2024
• Node.js optimized concurrent execution model for Vercel Functions
• Private beta customer @madeonverse reduced their compute cost by 50%
• Available today as opt-in public betahttps://t.co/iB3qvHiacz
New Releases
A new EventBridge metric was announced last week that measures the time between event ingestion and successful delivery to targets. This should be extremely useful for identifying retries and intermittent failures in large workflows.
Name change alert! AWS Application Composer is now AWS Infrastructure Composer. As far as I can tell, this is just a rebranding and functionally no different than before.
Last Words
What is your opinion on the sudden deprecation of a slew of AWS services? Do you think it’s a good thing? I’m sure those of us affected by a service deprecation have much stronger opinions than those of us who haven’t been affected… yet. Personally, I see it as a good thing. I hope this means there will be more focus on the core, highly-used services which will result in more features. But you never 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
This Week's Sponsor
