🦸 Serverless Superhero

Our serverless superhero this week is Damian Antonio Gitto Olguin, AWS Community Hero, AWS User Group Leader for UG Cordoba Argentina and Co-Founder at Teracloud. Damian is commonly seen at AWS events (I’ve seen him at every AWS event I have attended LOL), and he’s not only attending, he is always sharing his experiences as a speaker. Damian, it is always great to see you and I wanted to thank you for all that you do for the community.

💯 Spotlight

With cold starts now affecting your bill and not just the latency of your invocation, David Behroozi took another deep dive to find performance improvements for your Lambda functions. In his post, The Fastest Node 22 Lambda Coldstart Configuration, David explores different configurations Lambda functions that will help shave some time off of the Lambda cold starts. David is always taking a step further than most of us will to find optimizations that can make applications better and cheaper.

🔥 My Favorite Content

When choosing DynamoDB as the database for your application, there are some known tradeoffs made in order to have better scalability and infrastructure management. With the introduction of Amazon Aurora DSQL, the tradeoffs are less. In the first part of his series Amazon Aurora DSQL Sidecar to DynamoDB, Lee Gilmore walks us through a thorough example of using DSQL paired with DynamoDB to fill in the gaps. I suspect this will be a common pattern for people wanting to go back to relational database but don’t want to go through a full migration right now.

AI is slowly becoming (okay.. maybe not too slowly) something we interact with A LOT. But as with many things, this opens up the doors for many security vulnerabilities. If you have good automation pipelines you are probably already taking care of most of these, but let’s be honest, there are many applications out there with only the aspiration for automation pipelines but haven’t gotten there yet. Eyal Estrin shows us the type of vulnerabilities we might be opening ourselves to when vibe coding, but thankfully, he also provides recommendations for how to mitigate these in his latest post, Common security pitfalls using Vibe coding

I hadn’t heard about ATProto until Danielle Heberling brought it up in the Believe In Serverless community. At a super high level, ATProto is a decentralized, open-source social networking protocol that allows us to own our data, control our identity and move seamlessly across interoperable platforms. In this post, Inanna Malick gives a brief overview on what this protocol is and shows us how to build and deploy a serverless ATProto application in Cloudflare.

Whenever AWS releases something new, there are usually products out there that it competes with. The Lambda Live Debugging functionality is no different. Marko Štrukelj released the Lambda Live Debugger about a year, and this week he posted about the comparison between his solution and the one provided by AWS.

💡 Tip of the Week

I can’t stop thinking about Ultron trying to save the world by destroying it with this piece of information that Yan Cui shared about a PR that got submitted to Amazon Q by a hacker that could have been very problematic for our machines and infrastructure.
We are using AI tools constantly now, and sometimes the easiest path to get something to work is to give it access to run a wildcarded command on your machine (eg. git commit *). When giving an AI assistant access to run commands on your machine, be sure to give it restricted access to any that could be destructive or harmful. If you are unsure if it could be harmful, don’t give it permissions and have the agent wait for you to run the command instead. I know this can slow you down, but that’s the thing about security, if it isn’t inconvenient, is it even secure?

🐣 New Releases

After a jam packed week of releases at the AWS Summit in New York, this week was fairly slow.

Amazon SQS introduces fair queues for multi-tenant workloads. This is a great feature that reduces the risk of a customer being impacted by a different customer that is overloading the system. With this feature, SQS will automatically detect the “noisy customer” and prioritizes messages that belong to a “quiet customer”. This means that the queue backlog for noisy customers will grow without slowing down the quiet customers.

Last Words

This weekend I had the opportunity to go camping and I can’t explain how great it felt to disconnect from the internet and simply enjoy nature. It was so refreshing and I am back with more energy and motivation to get back to work. I hope you’ve enjoyed my summary of last weeks content, as always it’s been a pleasure to be able to share my thoughts with you.

Until next time

Andres