Tobias Schmidt
@tpschmidt.com
Helping aspiring engineers master the cloud
👨💻 Freelance Software Engineer
✍️ Book #1: http://awsfundamentals.com
📕 Book #2: http://cloudwatchbook.com
Learn AWS for Free: https://awsfundamentals.com/newsletter
👨💻 Freelance Software Engineer
✍️ Book #1: http://awsfundamentals.com
📕 Book #2: http://cloudwatchbook.com
Learn AWS for Free: https://awsfundamentals.com/newsletter
AWS increased the payload limit from 256 KB to 1 MB for Lambda async invocations, SQS, and EventBridge! 🎉
This is a bigger deal than it sounds!
Dumping payloads to S3 and passing references can be a hastle, even with helpers that support that ootb.
This is a bigger deal than it sounds!
Dumping payloads to S3 and passing references can be a hastle, even with helpers that support that ootb.
February 2, 2026 at 2:57 PM
AWS increased the payload limit from 256 KB to 1 MB for Lambda async invocations, SQS, and EventBridge! 🎉
This is a bigger deal than it sounds!
Dumping payloads to S3 and passing references can be a hastle, even with helpers that support that ootb.
This is a bigger deal than it sounds!
Dumping payloads to S3 and passing references can be a hastle, even with helpers that support that ootb.
Vantage built an EC2 instance comparison tool that's actually useful.
It's at instances.vantage.sh and beats the AWS console by a mile.
You can filter by region, compare pricing (on-demand, reserved, spot), and sort by actual performance benchmarks scores (and even FFmpeg FPS).
It's at instances.vantage.sh and beats the AWS console by a mile.
You can filter by region, compare pricing (on-demand, reserved, spot), and sort by actual performance benchmarks scores (and even FFmpeg FPS).
February 2, 2026 at 8:00 AM
Vantage built an EC2 instance comparison tool that's actually useful.
It's at instances.vantage.sh and beats the AWS console by a mile.
You can filter by region, compare pricing (on-demand, reserved, spot), and sort by actual performance benchmarks scores (and even FFmpeg FPS).
It's at instances.vantage.sh and beats the AWS console by a mile.
You can filter by region, compare pricing (on-demand, reserved, spot), and sort by actual performance benchmarks scores (and even FFmpeg FPS).
I've been using LocalStack for a while now and honestly wish I'd started sooner 🛠️
It's a fully functional AWS cloud stack that runs entirely on your machine.
It's a fully functional AWS cloud stack that runs entirely on your machine.
February 1, 2026 at 2:59 PM
I've been using LocalStack for a while now and honestly wish I'd started sooner 🛠️
It's a fully functional AWS cloud stack that runs entirely on your machine.
It's a fully functional AWS cloud stack that runs entirely on your machine.
One of the scariest things in AWS is a developer "accidentally" spinning up a 𝗽𝟰𝗱.𝟮𝟰𝘅𝗹𝗮𝗿𝗴𝗲 instance.
That's $22/h ($16k/m) gone if you forget about it 💸
You can try to catch this with budget alerts, but by the time you get the email, the money is already spent.
The better way?
That's $22/h ($16k/m) gone if you forget about it 💸
You can try to catch this with budget alerts, but by the time you get the email, the money is already spent.
The better way?
February 1, 2026 at 8:00 AM
One of the scariest things in AWS is a developer "accidentally" spinning up a 𝗽𝟰𝗱.𝟮𝟰𝘅𝗹𝗮𝗿𝗴𝗲 instance.
That's $22/h ($16k/m) gone if you forget about it 💸
You can try to catch this with budget alerts, but by the time you get the email, the money is already spent.
The better way?
That's $22/h ($16k/m) gone if you forget about it 💸
You can try to catch this with budget alerts, but by the time you get the email, the money is already spent.
The better way?
I've been routing API traffic through CloudFront just to do A/B tests for years. Turns out AWS quietly shipped a feature that makes this whole setup unnecessary.
January 31, 2026 at 8:01 AM
I've been routing API traffic through CloudFront just to do A/B tests for years. Turns out AWS quietly shipped a feature that makes this whole setup unnecessary.
Most Lambda functions I audit are burning 20-30% more than they should.
Not because of complex architecture.
Just basic config mistakes that never get fixed 🤷♂️
The quick wins:
Not because of complex architecture.
Just basic config mistakes that never get fixed 🤷♂️
The quick wins:
January 30, 2026 at 7:56 AM
Most Lambda functions I audit are burning 20-30% more than they should.
Not because of complex architecture.
Just basic config mistakes that never get fixed 🤷♂️
The quick wins:
Not because of complex architecture.
Just basic config mistakes that never get fixed 🤷♂️
The quick wins:
You likely know that EventBridge exists.
But do you know it's perfect for centralizing alerts across multiple accounts?
Here's the pattern we keep coming back to 👇
The problem with AWS alerts is they come from everywhere, so different services, accounts & regions.
But do you know it's perfect for centralizing alerts across multiple accounts?
Here's the pattern we keep coming back to 👇
The problem with AWS alerts is they come from everywhere, so different services, accounts & regions.
January 29, 2026 at 3:03 PM
You likely know that EventBridge exists.
But do you know it's perfect for centralizing alerts across multiple accounts?
Here's the pattern we keep coming back to 👇
The problem with AWS alerts is they come from everywhere, so different services, accounts & regions.
But do you know it's perfect for centralizing alerts across multiple accounts?
Here's the pattern we keep coming back to 👇
The problem with AWS alerts is they come from everywhere, so different services, accounts & regions.
Anton Babenko's Terraform Claude skill completely changed how I write infrastructure code. And no, it's not just "another AI" that knows Terraform syntax.
This thing actually follows battle-tested workflows and best practices.
If you missed the release, here's a small wrap up:
This thing actually follows battle-tested workflows and best practices.
If you missed the release, here's a small wrap up:
January 29, 2026 at 8:03 AM
Anton Babenko's Terraform Claude skill completely changed how I write infrastructure code. And no, it's not just "another AI" that knows Terraform syntax.
This thing actually follows battle-tested workflows and best practices.
If you missed the release, here's a small wrap up:
This thing actually follows battle-tested workflows and best practices.
If you missed the release, here's a small wrap up:
With Kiro & a few MCP servers, you can visualize your architecture in a pretty usable way. And if something changes, you can just regenerate it!
The workflow is simple.
You either:
1. let the LLM check your IaC
2. or use an MCP server to query the resources in your AWS account.
The workflow is simple.
You either:
1. let the LLM check your IaC
2. or use an MCP server to query the resources in your AWS account.
January 28, 2026 at 2:59 PM
With Kiro & a few MCP servers, you can visualize your architecture in a pretty usable way. And if something changes, you can just regenerate it!
The workflow is simple.
You either:
1. let the LLM check your IaC
2. or use an MCP server to query the resources in your AWS account.
The workflow is simple.
You either:
1. let the LLM check your IaC
2. or use an MCP server to query the resources in your AWS account.
I've written the same deployment script 6 times. Multiple ECS clusters. Proxied Route53 records. Weight-shifting logic that breaks. AWS finally killed all that custom code last year 💛
ECS finally has built-in blue/green deployments ✨
ECS finally has built-in blue/green deployments ✨
January 28, 2026 at 8:02 AM
I've written the same deployment script 6 times. Multiple ECS clusters. Proxied Route53 records. Weight-shifting logic that breaks. AWS finally killed all that custom code last year 💛
ECS finally has built-in blue/green deployments ✨
ECS finally has built-in blue/green deployments ✨
AI can write infrastructure code that looks right.
Battle-tested patterns that actually survive production? Not so much.
Anton Babenko's Claude skill bridges that gap.
What it covers:
🤖 The Engine:
Strict engineering loop (init, validate, plan) with automated formatting.
Battle-tested patterns that actually survive production? Not so much.
Anton Babenko's Claude skill bridges that gap.
What it covers:
🤖 The Engine:
Strict engineering loop (init, validate, plan) with automated formatting.
January 27, 2026 at 12:13 PM
AI can write infrastructure code that looks right.
Battle-tested patterns that actually survive production? Not so much.
Anton Babenko's Claude skill bridges that gap.
What it covers:
🤖 The Engine:
Strict engineering loop (init, validate, plan) with automated formatting.
Battle-tested patterns that actually survive production? Not so much.
Anton Babenko's Claude skill bridges that gap.
What it covers:
🤖 The Engine:
Strict engineering loop (init, validate, plan) with automated formatting.
There's no prize for using only Lambda.
Stop treating Serverless like a religion.
If it’s not Lambda, some people feel like they’re breaking the rules.
That mindset gets you stuck.
There’s no prize for using only _one set_ of AWS services.
Stop treating Serverless like a religion.
If it’s not Lambda, some people feel like they’re breaking the rules.
That mindset gets you stuck.
There’s no prize for using only _one set_ of AWS services.
January 27, 2026 at 8:02 AM
There's no prize for using only Lambda.
Stop treating Serverless like a religion.
If it’s not Lambda, some people feel like they’re breaking the rules.
That mindset gets you stuck.
There’s no prize for using only _one set_ of AWS services.
Stop treating Serverless like a religion.
If it’s not Lambda, some people feel like they’re breaking the rules.
That mindset gets you stuck.
There’s no prize for using only _one set_ of AWS services.
AWS has 300+ services.
You need maybe 20 of them.
I've shipped dozens of projects over the years.
The same core services show up in every single one.
Trying to learn everything is overwhelming.
So you jump between dozends of services, never going deep on anything.
That's backwards.
You need maybe 20 of them.
I've shipped dozens of projects over the years.
The same core services show up in every single one.
Trying to learn everything is overwhelming.
So you jump between dozends of services, never going deep on anything.
That's backwards.
January 26, 2026 at 2:57 PM
AWS has 300+ services.
You need maybe 20 of them.
I've shipped dozens of projects over the years.
The same core services show up in every single one.
Trying to learn everything is overwhelming.
So you jump between dozends of services, never going deep on anything.
That's backwards.
You need maybe 20 of them.
I've shipped dozens of projects over the years.
The same core services show up in every single one.
Trying to learn everything is overwhelming.
So you jump between dozends of services, never going deep on anything.
That's backwards.
Testing CloudFront and Lambda@Edge used to be painful.
You'd deploy, wait (>5m), test, realize something broke, then repeat the whole cycle 😅
That's why I'm using LocalStack 🏗️
You'd deploy, wait (>5m), test, realize something broke, then repeat the whole cycle 😅
That's why I'm using LocalStack 🏗️
January 26, 2026 at 8:03 AM
Testing CloudFront and Lambda@Edge used to be painful.
You'd deploy, wait (>5m), test, realize something broke, then repeat the whole cycle 😅
That's why I'm using LocalStack 🏗️
You'd deploy, wait (>5m), test, realize something broke, then repeat the whole cycle 😅
That's why I'm using LocalStack 🏗️
Most AWS tool lists are not great 😅
So I made one page with only tools I'm _actually_ using 📚
It includes:
🏗️ Awesome Terraform – Curated list of Terraform resources for Infrastructure as Code best practices and useful tools.
So I made one page with only tools I'm _actually_ using 📚
It includes:
🏗️ Awesome Terraform – Curated list of Terraform resources for Infrastructure as Code best practices and useful tools.
January 25, 2026 at 3:03 PM
Most AWS tool lists are not great 😅
So I made one page with only tools I'm _actually_ using 📚
It includes:
🏗️ Awesome Terraform – Curated list of Terraform resources for Infrastructure as Code best practices and useful tools.
So I made one page with only tools I'm _actually_ using 📚
It includes:
🏗️ Awesome Terraform – Curated list of Terraform resources for Infrastructure as Code best practices and useful tools.
The thing I value most about AWS isn't the service catalog or pricing.
It's that I can write code against their APIs and trust it'll still work in 5 years.
Sure, they deprecate services and 30+ in 2025 alone 😅
But if a service survives past general availability, the API contract _stays solid_
It's that I can write code against their APIs and trust it'll still work in 5 years.
Sure, they deprecate services and 30+ in 2025 alone 😅
But if a service survives past general availability, the API contract _stays solid_
January 25, 2026 at 8:00 AM
The thing I value most about AWS isn't the service catalog or pricing.
It's that I can write code against their APIs and trust it'll still work in 5 years.
Sure, they deprecate services and 30+ in 2025 alone 😅
But if a service survives past general availability, the API contract _stays solid_
It's that I can write code against their APIs and trust it'll still work in 5 years.
Sure, they deprecate services and 30+ in 2025 alone 😅
But if a service survives past general availability, the API contract _stays solid_
IAM isn't hard because you're bad at it.
It's hard because AWS made it complicated as hell 🤷
But here's what I see most people get wrong about roles.
They treat them like "advanced" permission sets.
Just another place to attach policies and move on.
That's missing the whole point.
It's hard because AWS made it complicated as hell 🤷
But here's what I see most people get wrong about roles.
They treat them like "advanced" permission sets.
Just another place to attach policies and move on.
That's missing the whole point.
January 24, 2026 at 3:03 PM
IAM isn't hard because you're bad at it.
It's hard because AWS made it complicated as hell 🤷
But here's what I see most people get wrong about roles.
They treat them like "advanced" permission sets.
Just another place to attach policies and move on.
That's missing the whole point.
It's hard because AWS made it complicated as hell 🤷
But here's what I see most people get wrong about roles.
They treat them like "advanced" permission sets.
Just another place to attach policies and move on.
That's missing the whole point.
Semantic search without a vector database sounds impossible.
Until you realize your data already lives in S3.
Pinecone and similar services work great but it's another service to maintain, monitor, and pay for separately.
S3 Vectors changed this.
Until you realize your data already lives in S3.
Pinecone and similar services work great but it's another service to maintain, monitor, and pay for separately.
S3 Vectors changed this.
January 24, 2026 at 7:58 AM
Semantic search without a vector database sounds impossible.
Until you realize your data already lives in S3.
Pinecone and similar services work great but it's another service to maintain, monitor, and pay for separately.
S3 Vectors changed this.
Until you realize your data already lives in S3.
Pinecone and similar services work great but it's another service to maintain, monitor, and pay for separately.
S3 Vectors changed this.
Most developers don't actually debug Lambda functions.
They just redeploy with more console.log statements until something works.
Last year, AWS finally fixed this with 𝗿𝗲𝗮𝗹 𝗿𝗲𝗺𝗼𝘁𝗲 𝗱𝗲𝗯𝘂𝗴𝗴𝗶𝗻𝗴 of deployed functions.
Why is this a big deal?
They just redeploy with more console.log statements until something works.
Last year, AWS finally fixed this with 𝗿𝗲𝗮𝗹 𝗿𝗲𝗺𝗼𝘁𝗲 𝗱𝗲𝗯𝘂𝗴𝗴𝗶𝗻𝗴 of deployed functions.
Why is this a big deal?
January 23, 2026 at 2:59 PM
Most developers don't actually debug Lambda functions.
They just redeploy with more console.log statements until something works.
Last year, AWS finally fixed this with 𝗿𝗲𝗮𝗹 𝗿𝗲𝗺𝗼𝘁𝗲 𝗱𝗲𝗯𝘂𝗴𝗴𝗶𝗻𝗴 of deployed functions.
Why is this a big deal?
They just redeploy with more console.log statements until something works.
Last year, AWS finally fixed this with 𝗿𝗲𝗮𝗹 𝗿𝗲𝗺𝗼𝘁𝗲 𝗱𝗲𝗯𝘂𝗴𝗴𝗶𝗻𝗴 of deployed functions.
Why is this a big deal?
You probably don't need a dedicated vector database.
S3 Vectors made my favorite AWS releases last year by proving it 💪
It's native vector storage built right into S3. No infrastructure to manage. Sub-second queries. Up to 90% cheaper than running dedicated vector databases.
S3 Vectors made my favorite AWS releases last year by proving it 💪
It's native vector storage built right into S3. No infrastructure to manage. Sub-second queries. Up to 90% cheaper than running dedicated vector databases.
January 23, 2026 at 8:02 AM
You probably don't need a dedicated vector database.
S3 Vectors made my favorite AWS releases last year by proving it 💪
It's native vector storage built right into S3. No infrastructure to manage. Sub-second queries. Up to 90% cheaper than running dedicated vector databases.
S3 Vectors made my favorite AWS releases last year by proving it 💪
It's native vector storage built right into S3. No infrastructure to manage. Sub-second queries. Up to 90% cheaper than running dedicated vector databases.
I've spent 6 years answering the same AWS questions over and over.
So we compiled dozens of practical tip into one page 💁♂️
Just the stuff that actually saves you time and money - or maybe confused your heavily - when working with AWS.
We're adding more every week! ✨
So we compiled dozens of practical tip into one page 💁♂️
Just the stuff that actually saves you time and money - or maybe confused your heavily - when working with AWS.
We're adding more every week! ✨
January 22, 2026 at 2:59 PM
I've spent 6 years answering the same AWS questions over and over.
So we compiled dozens of practical tip into one page 💁♂️
Just the stuff that actually saves you time and money - or maybe confused your heavily - when working with AWS.
We're adding more every week! ✨
So we compiled dozens of practical tip into one page 💁♂️
Just the stuff that actually saves you time and money - or maybe confused your heavily - when working with AWS.
We're adding more every week! ✨
This meme hits different when you know the actual numbers.
When us-east-1 goes down, half the internet goes with it.
Not too much of an exaggeration 😅
The funny part? AWS keeps telling everyone to build multi region architectures. Meanwhile us-east-1 remains the backbone of the internet.
When us-east-1 goes down, half the internet goes with it.
Not too much of an exaggeration 😅
The funny part? AWS keeps telling everyone to build multi region architectures. Meanwhile us-east-1 remains the backbone of the internet.
January 22, 2026 at 8:00 AM
This meme hits different when you know the actual numbers.
When us-east-1 goes down, half the internet goes with it.
Not too much of an exaggeration 😅
The funny part? AWS keeps telling everyone to build multi region architectures. Meanwhile us-east-1 remains the backbone of the internet.
When us-east-1 goes down, half the internet goes with it.
Not too much of an exaggeration 😅
The funny part? AWS keeps telling everyone to build multi region architectures. Meanwhile us-east-1 remains the backbone of the internet.
I've never seen anyone learn AWS faster from reading documentation.
Visual flows, one-page summaries, and hands-on code?
That's what actually works.
We spent years building exactly that.
All free 📚
Here's what you'll find...
📊 𝗔𝗪𝗦 𝗜𝗻𝗳𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰𝘀
Visual flows, one-page summaries, and hands-on code?
That's what actually works.
We spent years building exactly that.
All free 📚
Here's what you'll find...
📊 𝗔𝗪𝗦 𝗜𝗻𝗳𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰𝘀
January 21, 2026 at 3:04 PM
I've never seen anyone learn AWS faster from reading documentation.
Visual flows, one-page summaries, and hands-on code?
That's what actually works.
We spent years building exactly that.
All free 📚
Here's what you'll find...
📊 𝗔𝗪𝗦 𝗜𝗻𝗳𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰𝘀
Visual flows, one-page summaries, and hands-on code?
That's what actually works.
We spent years building exactly that.
All free 📚
Here's what you'll find...
📊 𝗔𝗪𝗦 𝗜𝗻𝗳𝗼𝗴𝗿𝗮𝗽𝗵𝗶𝗰𝘀
We're overwhelmed by the support for the launch of our YouTube channel.
We're close to reaching the first 1k subscribers! 🎉 🥳
If you want to join, we're happy to have you!
What to expect?
We want to help developers bridge the gap between certification knowledge and real-world AWS skills.
We're close to reaching the first 1k subscribers! 🎉 🥳
If you want to join, we're happy to have you!
What to expect?
We want to help developers bridge the gap between certification knowledge and real-world AWS skills.
January 21, 2026 at 8:04 AM
We're overwhelmed by the support for the launch of our YouTube channel.
We're close to reaching the first 1k subscribers! 🎉 🥳
If you want to join, we're happy to have you!
What to expect?
We want to help developers bridge the gap between certification knowledge and real-world AWS skills.
We're close to reaching the first 1k subscribers! 🎉 🥳
If you want to join, we're happy to have you!
What to expect?
We want to help developers bridge the gap between certification knowledge and real-world AWS skills.
The best Lambda debugging feature AWS ever shipped and I rarely see anyone using it. Wild because it solves the exact problem every serverless dev complains about.
Talking about remote debugging of deployed Lambda functions!
Talking about remote debugging of deployed Lambda functions!
January 20, 2026 at 3:01 PM
The best Lambda debugging feature AWS ever shipped and I rarely see anyone using it. Wild because it solves the exact problem every serverless dev complains about.
Talking about remote debugging of deployed Lambda functions!
Talking about remote debugging of deployed Lambda functions!