Tobias Schmidt

With AWS worsening their news page regularly, here's likely the slickest "I want to stay up with everything" page out there! ✨ No weird formatting or controls. Just a simple list with dates and links. Love it!

zerowastecloud.io/aws-whats-new

Sometimes it’s uncomfortable.
Sometimes it’s exactly what I need.

If you’re building something, don’t underestimate the random connections. The best stuff usually starts with a DM or a tweet you almost didn’t send.

Funny thing: somewhere along the way, Sandro went from random internet guy to one of my closest friends. I’ve learned more from him than from any book or course.

Every strong opinion gets challenged.
Neither of us lets stuff slide, and it’s forced both of us to level up.

Fast forward, we bootstrapped AWS Fundamentals together.
No outside money, just a lot of late nights, way too much coffee, and more arguments than I can count.

We don’t always agree - actually, we rarely do at first.
That’s probably why this works.

Wild how things start. A couple years back I fired off a random tweet looking for folks also building things on AWS.
Out of nowhere, Sandro replied.

We had zero plan. Just two strangers, both obsessed with building.

Amazon SQS increases maximum message payload size to 1 MiB - AWS Discover more about what's new at AWS with Amazon SQS increases maximum message payload size to 1 MiB

Link to announcement 🔗
aws.amazon.com/about-aws/w...

If you’ve built anything that pushes real data through SQS, you know the old limit forced some awkward workarounds.

Either chunk the message and reassemble downstream, or push the real payload to S3 and just pass a pointer. Nobody wants to deal with that. 😅

If you missed it: SQS (finally) bumped the max message size from 256 KiB to 1 MiB 📈
Would be happy to see this too at Azure's ServiceBus for the non-premium tier 😪

Reminder: you're still charged in 64kb chunks 💸

AWS Lambda response streaming now supports 200 MB response payloads - AWS Discover more about what's new at AWS with AWS Lambda response streaming now supports 200 MB response payloads

Official news post by AWS 🔗
aws.amazon.com/about-aws/w...

Hit the 6MB Lambda payload limit? Here's what you can do. Having trouble with the 6MB payload limit for AWS Lambda? Explore four solutions to overcome this limitation.

Second disclaimer: still sitting there with the 6MB input limit 😬
But there are ways around that:
theburningmonk.com/2020/04/hit...

𝗗𝗶𝘀𝗰𝗹𝗮𝗶𝗺𝗲𝗿: this is for response streaming, not the classic Lambda invoke. And you’ll need to use Node.js managed runtimes or a custom runtime.

200 MB isn’t infinite, but it’s enough to cover most “big response” use cases without using other infrastructure.

Now you can stream up to 200 MB straight from Lambda. No more S3 handoff for every oversized response.

Before, if you needed to send more than 20 MB, you had to compress, chunk, or offload to S3 via presigned URLs. More moving parts, more latency (and more code to maintain! Problably the worst trade-off before).

Lambda response streaming now handles 200 MB payloads by default. That’s 10x the old limit.

For any workload where you’re pushing big responses, mostly large media, this simplifies your infrastructure (and code)! 🤝

From the aws community on Reddit Explore this post and more from the aws community

Here's the link to the discussion.
Great and diverse answers.

www.reddit.com/r/aws/comme...

I think this is the same as with observability.
It's not a nice-to-have that magically runs on the side.
Everybody needs to be involved and it's a critical part of every software.

I just read a discussion on Reddit about how to get developers to care about AWS costs 💸.
The top answer is spot-on: cost ownership can't be centralized.
It needs to be part of every team.

A team needs to own its stack from top to bottom.
Including its costs.

Please do not create shared resources between different stages.
At best, you'll have dedicated accounts per stage. 👥
Take advantage of the natural security boundaries for an AWS account.

Probably you can't hear it anymore but: yes, please use AWS Organizations too 🤓

TIL: even without using Organization's centralized billing, your bank statement won't show which account AWS charged you for 💭

Nevertheless, the major takeaway remains: please use AWS Organization and centralized billing and take a look at your cost reports! 💸

If a message skips an attribute your filter cares about, you don’t get it.

TL;DR:
• Each message can have up to 10 attributes.
• Filters are JSON policies.
• Matching supports strings, numbers, “anything-but”, and prefix.

SNS now lets you attach a filter policy to each subscription, so subscribers only get the messages they actually care about.

Each message will be checked against each subscription’s filter, and only delivers if there’s a match.

Not a small detail if you’re running experiments or onboarding new projects.

AWS docs mention this, but it’s buried. If you’re running training environments, POCs, or dev/test workloads relying on Free Tier - be aware of this!

Anyone else already got burned by this? 😅

You also can’t earn up to $100 more in Free Tier credits.

This catches a lot of ambitious learners off guard, I think. You spin up an org to get serious about account management, security, or billing, and suddenly your credits are gone.

𝗧𝗜𝗟: The AWS Free Tier isn’t free anymore once you start using Organizations or Control Tower! 💸

Join an AWS Organization or set up a Control Tower landing zone, and your Free Tier credits 𝗱𝗶𝘀𝗮𝗽𝗽𝗲𝗮𝗿. Instantly. No warning, no grace period. You’re now on a paid plan.

Another cheat sheet from the past: this one was for the Lambda CLI λ
Today, I'd recommend the VSCode extension or better yet, use something like SST to have the best Lambda developer experience! ⚡️

Introducing the Amazon DynamoDB data modeling MCP tool | Amazon Web Services To help you move faster with greater confidence, we’re introducing a new DynamoDB data modeling tool, available as part of our DynamoDB Model Context Protocol (MCP) server. The DynamoDB MCP data modeling tool integrates with AI assistants that support MCP, providing a structured, natural-language-driven workflow to translate application requirements into DynamoDB data models. In this post, we show you how to generate a data model in minutes using this new data modeling tool.

Link to the announcement article 🔗
aws.amazon.com/blogs/datab...

If you’re tired of “wizard-driven” tools that ignore how your app actually works, this is a step in the right direction.

Looks super promising.

You can review them, iterate, and hopefully understand the why behind each choice.

Once you’re done, it generates out a 𝗿𝗲𝗾𝘂𝗶𝗿𝗲𝗺𝗲𝗻𝘁𝘀.𝗺𝗱 and a 𝗱𝗮𝘁𝗮_𝗺𝗼𝗱𝗲𝗹.𝗺𝗱, including tables, indexes, reasoning, cost estimates, even trade-offs.