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
I thought I understood how Lambda scales under the hood.
Then I read about Shuffle Sharding...
It turns out, preventing noisy neighbors isn't just about adding more servers. It's about math.
AWS just shared how Lambda handles billions of asynchronous invocations.
Then I read about Shuffle Sharding...
It turns out, preventing noisy neighbors isn't just about adding more servers. It's about math.
AWS just shared how Lambda handles billions of asynchronous invocations.
December 27, 2025 at 2:56 PM
I thought I understood how Lambda scales under the hood.
Then I read about Shuffle Sharding...
It turns out, preventing noisy neighbors isn't just about adding more servers. It's about math.
AWS just shared how Lambda handles billions of asynchronous invocations.
Then I read about Shuffle Sharding...
It turns out, preventing noisy neighbors isn't just about adding more servers. It's about math.
AWS just shared how Lambda handles billions of asynchronous invocations.
Giving developers permanent Admin access is mostly a terrible idea.
Much worse: blocking them from fixing issues on prod.
I've already promoted 𝗧𝗘𝗔𝗠 (Temporary Elevated Access Management) a while ago, but it's just great and it solves this exact problem!
Much worse: blocking them from fixing issues on prod.
I've already promoted 𝗧𝗘𝗔𝗠 (Temporary Elevated Access Management) a while ago, but it's just great and it solves this exact problem!
December 27, 2025 at 7:58 AM
Giving developers permanent Admin access is mostly a terrible idea.
Much worse: blocking them from fixing issues on prod.
I've already promoted 𝗧𝗘𝗔𝗠 (Temporary Elevated Access Management) a while ago, but it's just great and it solves this exact problem!
Much worse: blocking them from fixing issues on prod.
I've already promoted 𝗧𝗘𝗔𝗠 (Temporary Elevated Access Management) a while ago, but it's just great and it solves this exact problem!
Sometimes you can cut your AWS bill in half with just an hour of work! 📉
It’s an easy way to shine at your company.
I recently looked at an account where we did exactly that.
The changes were minimal, but the impact was huge.
Here is what we fixed
1️⃣ 𝗠𝗲𝘁𝗿𝗶𝗰 𝘀𝗮𝗺𝗽𝗹𝗲 𝗿𝗮𝘁𝗲𝘀
It’s an easy way to shine at your company.
I recently looked at an account where we did exactly that.
The changes were minimal, but the impact was huge.
Here is what we fixed
1️⃣ 𝗠𝗲𝘁𝗿𝗶𝗰 𝘀𝗮𝗺𝗽𝗹𝗲 𝗿𝗮𝘁𝗲𝘀
December 26, 2025 at 3:04 PM
Sometimes you can cut your AWS bill in half with just an hour of work! 📉
It’s an easy way to shine at your company.
I recently looked at an account where we did exactly that.
The changes were minimal, but the impact was huge.
Here is what we fixed
1️⃣ 𝗠𝗲𝘁𝗿𝗶𝗰 𝘀𝗮𝗺𝗽𝗹𝗲 𝗿𝗮𝘁𝗲𝘀
It’s an easy way to shine at your company.
I recently looked at an account where we did exactly that.
The changes were minimal, but the impact was huge.
Here is what we fixed
1️⃣ 𝗠𝗲𝘁𝗿𝗶𝗰 𝘀𝗮𝗺𝗽𝗹𝗲 𝗿𝗮𝘁𝗲𝘀
I hate leaving the terminal just to check my current AWS bill.
Nugget I found last year: 𝗔𝗪𝗦 𝗙𝗶𝗻𝗢𝗽𝘀 𝗗𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 ✨
It visualizes your cost breakdown right in the CLI.
Fast, simple, and actually looks good.
Open source and built by Ravi!
Link to the repo 👇
Nugget I found last year: 𝗔𝗪𝗦 𝗙𝗶𝗻𝗢𝗽𝘀 𝗗𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 ✨
It visualizes your cost breakdown right in the CLI.
Fast, simple, and actually looks good.
Open source and built by Ravi!
Link to the repo 👇
December 26, 2025 at 7:59 AM
I hate leaving the terminal just to check my current AWS bill.
Nugget I found last year: 𝗔𝗪𝗦 𝗙𝗶𝗻𝗢𝗽𝘀 𝗗𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 ✨
It visualizes your cost breakdown right in the CLI.
Fast, simple, and actually looks good.
Open source and built by Ravi!
Link to the repo 👇
Nugget I found last year: 𝗔𝗪𝗦 𝗙𝗶𝗻𝗢𝗽𝘀 𝗗𝗮𝘀𝗵𝗯𝗼𝗮𝗿𝗱 ✨
It visualizes your cost breakdown right in the CLI.
Fast, simple, and actually looks good.
Open source and built by Ravi!
Link to the repo 👇
Manual diagramming tools are the worst.
I literally spend hours aligning boxes and arrows.
Mostly it's okay to draw the first draft, especially with tools like Excalidraw, which are really simple to use, but when you need to make updates, it always gets painful.
I literally spend hours aligning boxes and arrows.
Mostly it's okay to draw the first draft, especially with tools like Excalidraw, which are really simple to use, but when you need to make updates, it always gets painful.
December 25, 2025 at 2:57 PM
Manual diagramming tools are the worst.
I literally spend hours aligning boxes and arrows.
Mostly it's okay to draw the first draft, especially with tools like Excalidraw, which are really simple to use, but when you need to make updates, it always gets painful.
I literally spend hours aligning boxes and arrows.
Mostly it's okay to draw the first draft, especially with tools like Excalidraw, which are really simple to use, but when you need to make updates, it always gets painful.
I recently picked up a cost optimization tip, and it worked out perfectly:
Moving our compute for dev stages to ap-south-1 🇮🇳
The reason is simple.
1. Spot prices are among the cheapest on the market there
Moving our compute for dev stages to ap-south-1 🇮🇳
The reason is simple.
1. Spot prices are among the cheapest on the market there
December 25, 2025 at 7:57 AM
I recently picked up a cost optimization tip, and it worked out perfectly:
Moving our compute for dev stages to ap-south-1 🇮🇳
The reason is simple.
1. Spot prices are among the cheapest on the market there
Moving our compute for dev stages to ap-south-1 🇮🇳
The reason is simple.
1. Spot prices are among the cheapest on the market there
It pays off to think about 𝗦𝗲𝗿𝘃𝗶𝗰𝗲 𝗖𝗼𝗻𝘁𝗿𝗼𝗹 𝗣𝗼𝗹𝗶𝗰𝗶𝗲𝘀 early on!
You can effectively stop bad habits before they even start.
Here are the ones I usually recommend to set up immediately:
1️⃣ 𝗕𝗹𝗼𝗰𝗸 𝗜𝗔𝗠 𝗨𝘀𝗲𝗿𝘀
You can effectively stop bad habits before they even start.
Here are the ones I usually recommend to set up immediately:
1️⃣ 𝗕𝗹𝗼𝗰𝗸 𝗜𝗔𝗠 𝗨𝘀𝗲𝗿𝘀
December 24, 2025 at 2:59 PM
It pays off to think about 𝗦𝗲𝗿𝘃𝗶𝗰𝗲 𝗖𝗼𝗻𝘁𝗿𝗼𝗹 𝗣𝗼𝗹𝗶𝗰𝗶𝗲𝘀 early on!
You can effectively stop bad habits before they even start.
Here are the ones I usually recommend to set up immediately:
1️⃣ 𝗕𝗹𝗼𝗰𝗸 𝗜𝗔𝗠 𝗨𝘀𝗲𝗿𝘀
You can effectively stop bad habits before they even start.
Here are the ones I usually recommend to set up immediately:
1️⃣ 𝗕𝗹𝗼𝗰𝗸 𝗜𝗔𝗠 𝗨𝘀𝗲𝗿𝘀
I knew us-east-1 was the "unreliable" region.
Didn't expect it to be nearly 𝟯𝘅 worse than the runner-up 😅
The numbers for 2025 are pretty clear:
- 10 major outages
- Almost 34 hours of total downtime
- 126 affected components
Didn't expect it to be nearly 𝟯𝘅 worse than the runner-up 😅
The numbers for 2025 are pretty clear:
- 10 major outages
- Almost 34 hours of total downtime
- 126 affected components
December 24, 2025 at 10:40 AM
I knew us-east-1 was the "unreliable" region.
Didn't expect it to be nearly 𝟯𝘅 worse than the runner-up 😅
The numbers for 2025 are pretty clear:
- 10 major outages
- Almost 34 hours of total downtime
- 126 affected components
Didn't expect it to be nearly 𝟯𝘅 worse than the runner-up 😅
The numbers for 2025 are pretty clear:
- 10 major outages
- Almost 34 hours of total downtime
- 126 affected components
Extended my "Cost Optimization Quick Fix" list after receiving many comments on the previous list! Now covers even more small things that can be easily implemented but can make a huge difference in your bill 💸
December 24, 2025 at 8:00 AM
Extended my "Cost Optimization Quick Fix" list after receiving many comments on the previous list! Now covers even more small things that can be easily implemented but can make a huge difference in your bill 💸
I used to avoid Lambda@Edge solely because of the developer experience.
Waiting for a CloudFront distribution to update just to test one header change? That workflow is finally fixed...
...with LocalStack 💪
You can now create distributions in milliseconds.
Waiting for a CloudFront distribution to update just to test one header change? That workflow is finally fixed...
...with LocalStack 💪
You can now create distributions in milliseconds.
December 23, 2025 at 2:58 PM
I used to avoid Lambda@Edge solely because of the developer experience.
Waiting for a CloudFront distribution to update just to test one header change? That workflow is finally fixed...
...with LocalStack 💪
You can now create distributions in milliseconds.
Waiting for a CloudFront distribution to update just to test one header change? That workflow is finally fixed...
...with LocalStack 💪
You can now create distributions in milliseconds.
In case you missed it: Lambda response streaming limit is now 200 MB.
That's a 10x increase from the previous 20 MB cap.
If you've been using S3 presigned URLs just to return a large PDF or image (and you can't use a CDN 👋), you can likely delete that code now.
That's a 10x increase from the previous 20 MB cap.
If you've been using S3 presigned URLs just to return a large PDF or image (and you can't use a CDN 👋), you can likely delete that code now.
December 23, 2025 at 2:31 PM
In case you missed it: Lambda response streaming limit is now 200 MB.
That's a 10x increase from the previous 20 MB cap.
If you've been using S3 presigned URLs just to return a large PDF or image (and you can't use a CDN 👋), you can likely delete that code now.
That's a 10x increase from the previous 20 MB cap.
If you've been using S3 presigned URLs just to return a large PDF or image (and you can't use a CDN 👋), you can likely delete that code now.
Almost missed this 🤯
There's a unified AWS MCP Server that covers all the small AWS MCP Servers that I already love! 🥳
Link to the docs ✏️
docs.aws.amazon.com/aws-mcp/lat...
There's a unified AWS MCP Server that covers all the small AWS MCP Servers that I already love! 🥳
Link to the docs ✏️
docs.aws.amazon.com/aws-mcp/lat...
December 23, 2025 at 8:02 AM
Almost missed this 🤯
There's a unified AWS MCP Server that covers all the small AWS MCP Servers that I already love! 🥳
Link to the docs ✏️
docs.aws.amazon.com/aws-mcp/lat...
There's a unified AWS MCP Server that covers all the small AWS MCP Servers that I already love! 🥳
Link to the docs ✏️
docs.aws.amazon.com/aws-mcp/lat...
The biggest blocker for AWS beginners isn't VPCs or IAM.
It's the fear of waking up to a $5,000 bill.
I speak to so many people who are terrified of the "cloud bill horror stories."
You wake up, check your account, and owe a 💩 of money because of a mess-up or a DDoS attack.
It's the fear of waking up to a $5,000 bill.
I speak to so many people who are terrified of the "cloud bill horror stories."
You wake up, check your account, and owe a 💩 of money because of a mess-up or a DDoS attack.
December 22, 2025 at 3:02 PM
The biggest blocker for AWS beginners isn't VPCs or IAM.
It's the fear of waking up to a $5,000 bill.
I speak to so many people who are terrified of the "cloud bill horror stories."
You wake up, check your account, and owe a 💩 of money because of a mess-up or a DDoS attack.
It's the fear of waking up to a $5,000 bill.
I speak to so many people who are terrified of the "cloud bill horror stories."
You wake up, check your account, and owe a 💩 of money because of a mess-up or a DDoS attack.
If you want to get into GenAI:
AWS offers a huge stack of free content on Skill Builder! 📚
I checked today: there are exactly 100 free items available!
Perfect for getting your hands dirty without the commitment.
Link is below 👇
AWS offers a huge stack of free content on Skill Builder! 📚
I checked today: there are exactly 100 free items available!
Perfect for getting your hands dirty without the commitment.
Link is below 👇
December 22, 2025 at 2:02 PM
If you want to get into GenAI:
AWS offers a huge stack of free content on Skill Builder! 📚
I checked today: there are exactly 100 free items available!
Perfect for getting your hands dirty without the commitment.
Link is below 👇
AWS offers a huge stack of free content on Skill Builder! 📚
I checked today: there are exactly 100 free items available!
Perfect for getting your hands dirty without the commitment.
Link is below 👇
EC2 Instance Savings Plans offer the highest discounts: up to 72%
But they come with the 𝗙𝗮𝗺𝗶𝗹𝘆 𝗧𝗿𝗮𝗽 that you should know about 👀
Looks great on paper: commit to a specific instance family (e.g., m5) in a specific region, and get massive savings 💸
The problem?
But they come with the 𝗙𝗮𝗺𝗶𝗹𝘆 𝗧𝗿𝗮𝗽 that you should know about 👀
Looks great on paper: commit to a specific instance family (e.g., m5) in a specific region, and get massive savings 💸
The problem?
December 22, 2025 at 7:58 AM
EC2 Instance Savings Plans offer the highest discounts: up to 72%
But they come with the 𝗙𝗮𝗺𝗶𝗹𝘆 𝗧𝗿𝗮𝗽 that you should know about 👀
Looks great on paper: commit to a specific instance family (e.g., m5) in a specific region, and get massive savings 💸
The problem?
But they come with the 𝗙𝗮𝗺𝗶𝗹𝘆 𝗧𝗿𝗮𝗽 that you should know about 👀
Looks great on paper: commit to a specific instance family (e.g., m5) in a specific region, and get massive savings 💸
The problem?
One of the biggest mistakes I see early AWS adopters make?
Waiting to set up AWS Organizations & Identity Center.
It feels like "enterprise stuff" when you're just starting out.
But fixing an awful account structure later is painful.
Setting it up early gives you a massive head start:
Waiting to set up AWS Organizations & Identity Center.
It feels like "enterprise stuff" when you're just starting out.
But fixing an awful account structure later is painful.
Setting it up early gives you a massive head start:
December 21, 2025 at 12:01 PM
One of the biggest mistakes I see early AWS adopters make?
Waiting to set up AWS Organizations & Identity Center.
It feels like "enterprise stuff" when you're just starting out.
But fixing an awful account structure later is painful.
Setting it up early gives you a massive head start:
Waiting to set up AWS Organizations & Identity Center.
It feels like "enterprise stuff" when you're just starting out.
But fixing an awful account structure later is painful.
Setting it up early gives you a massive head start:
You won't win any "Advanced Engineering" prizes for choosing ECS.
It's not Kubernetes. It's not shiny.
If you want to impress someone with complexity, pick something else.
But if you want to ship? Pick ECS.
It's the definition of 𝗯𝗮𝘁𝘁𝗹𝗲-𝘁𝗲𝘀𝘁𝗲𝗱.
It's not Kubernetes. It's not shiny.
If you want to impress someone with complexity, pick something else.
But if you want to ship? Pick ECS.
It's the definition of 𝗯𝗮𝘁𝘁𝗹𝗲-𝘁𝗲𝘀𝘁𝗲𝗱.
December 21, 2025 at 7:59 AM
You won't win any "Advanced Engineering" prizes for choosing ECS.
It's not Kubernetes. It's not shiny.
If you want to impress someone with complexity, pick something else.
But if you want to ship? Pick ECS.
It's the definition of 𝗯𝗮𝘁𝘁𝗹𝗲-𝘁𝗲𝘀𝘁𝗲𝗱.
It's not Kubernetes. It's not shiny.
If you want to impress someone with complexity, pick something else.
But if you want to ship? Pick ECS.
It's the definition of 𝗯𝗮𝘁𝘁𝗹𝗲-𝘁𝗲𝘀𝘁𝗲𝗱.
The "standard" AWS Tech Stack I find myself explaining to people again and again ✨
The services that actually matter for 90% of projects:
𝟭. 𝗖𝗼𝗺𝗽𝘂𝘁𝗲
The services that actually matter for 90% of projects:
𝟭. 𝗖𝗼𝗺𝗽𝘂𝘁𝗲
December 20, 2025 at 11:59 AM
The "standard" AWS Tech Stack I find myself explaining to people again and again ✨
The services that actually matter for 90% of projects:
𝟭. 𝗖𝗼𝗺𝗽𝘂𝘁𝗲
The services that actually matter for 90% of projects:
𝟭. 𝗖𝗼𝗺𝗽𝘂𝘁𝗲
If there is one service/feature that quietly glues the entire AWS ecosystem together, it's 𝗦𝗦𝗠 𝗣𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿 𝗦𝘁𝗼𝗿𝗲.
It's the silent powerhouse 💪
We often rush to complex solutions for configuration management.
But look at what this "boring" service actually handles for you:
It's the silent powerhouse 💪
We often rush to complex solutions for configuration management.
But look at what this "boring" service actually handles for you:
December 20, 2025 at 7:56 AM
If there is one service/feature that quietly glues the entire AWS ecosystem together, it's 𝗦𝗦𝗠 𝗣𝗮𝗿𝗮𝗺𝗲𝘁𝗲𝗿 𝗦𝘁𝗼𝗿𝗲.
It's the silent powerhouse 💪
We often rush to complex solutions for configuration management.
But look at what this "boring" service actually handles for you:
It's the silent powerhouse 💪
We often rush to complex solutions for configuration management.
But look at what this "boring" service actually handles for you:
I've just learned how Savings Plans are actually applied by the AWS billing engine.
It's smarter than I thought! 🧠
I always assumed it just applied the discount to whatever usage it saw first, or maybe randomly distributed it.
It's smarter than I thought! 🧠
I always assumed it just applied the discount to whatever usage it saw first, or maybe randomly distributed it.
December 20, 2025 at 7:02 AM
I've just learned how Savings Plans are actually applied by the AWS billing engine.
It's smarter than I thought! 🧠
I always assumed it just applied the discount to whatever usage it saw first, or maybe randomly distributed it.
It's smarter than I thought! 🧠
I always assumed it just applied the discount to whatever usage it saw first, or maybe randomly distributed it.
Did you know that CloudWatch Logs Insights supports query generation via LLMs for both OpenSearch PPL and SQL
And it works awesomely well! 💪
PPL doesn't have the most intuitive, so "how do I find X in these logs?" is a regular question.
And it works awesomely well! 💪
PPL doesn't have the most intuitive, so "how do I find X in these logs?" is a regular question.
December 19, 2025 at 2:01 PM
Did you know that CloudWatch Logs Insights supports query generation via LLMs for both OpenSearch PPL and SQL
And it works awesomely well! 💪
PPL doesn't have the most intuitive, so "how do I find X in these logs?" is a regular question.
And it works awesomely well! 💪
PPL doesn't have the most intuitive, so "how do I find X in these logs?" is a regular question.
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?
December 19, 2025 at 11:58 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?
My end-of-year AWS account cleanup checklist 🎄
December is when I take a few hours to audit my AWS organization and look for waste. This year I found some expensive surprises 😅
Here's what I check:
𝗖𝗼𝘀𝘁 𝗥𝗲𝗽𝗼𝗿𝘁𝘀 𝗮𝗻𝗱 𝗧𝗿𝗲𝗻𝗱𝘀
December is when I take a few hours to audit my AWS organization and look for waste. This year I found some expensive surprises 😅
Here's what I check:
𝗖𝗼𝘀𝘁 𝗥𝗲𝗽𝗼𝗿𝘁𝘀 𝗮𝗻𝗱 𝗧𝗿𝗲𝗻𝗱𝘀
December 19, 2025 at 8:00 AM
My end-of-year AWS account cleanup checklist 🎄
December is when I take a few hours to audit my AWS organization and look for waste. This year I found some expensive surprises 😅
Here's what I check:
𝗖𝗼𝘀𝘁 𝗥𝗲𝗽𝗼𝗿𝘁𝘀 𝗮𝗻𝗱 𝗧𝗿𝗲𝗻𝗱𝘀
December is when I take a few hours to audit my AWS organization and look for waste. This year I found some expensive surprises 😅
Here's what I check:
𝗖𝗼𝘀𝘁 𝗥𝗲𝗽𝗼𝗿𝘁𝘀 𝗮𝗻𝗱 𝗧𝗿𝗲𝗻𝗱𝘀
Finally: We have Database Savings Plans! This was the missing piece for a long time.
But before you purchase: understand the trade-offs!
AWS pricing is a wild jungle, and it's easy to get lost.
I wrote a guide to help you navigate the three main paths:
But before you purchase: understand the trade-offs!
AWS pricing is a wild jungle, and it's easy to get lost.
I wrote a guide to help you navigate the three main paths:
December 19, 2025 at 12:01 AM
Finally: We have Database Savings Plans! This was the missing piece for a long time.
But before you purchase: understand the trade-offs!
AWS pricing is a wild jungle, and it's easy to get lost.
I wrote a guide to help you navigate the three main paths:
But before you purchase: understand the trade-offs!
AWS pricing is a wild jungle, and it's easy to get lost.
I wrote a guide to help you navigate the three main paths:
"How efficient are we using AWS?" I guess most devs never thought about this, or they just (don't need to) care about it at all.
AWS can now give you a straight number!
It's called the 𝗖𝗼𝘀𝘁 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆 𝗦𝗰𝗼𝗿𝗲 👀
AWS can now give you a straight number!
It's called the 𝗖𝗼𝘀𝘁 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆 𝗦𝗰𝗼𝗿𝗲 👀
December 18, 2025 at 2:02 PM
"How efficient are we using AWS?" I guess most devs never thought about this, or they just (don't need to) care about it at all.
AWS can now give you a straight number!
It's called the 𝗖𝗼𝘀𝘁 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆 𝗦𝗰𝗼𝗿𝗲 👀
AWS can now give you a straight number!
It's called the 𝗖𝗼𝘀𝘁 𝗘𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆 𝗦𝗰𝗼𝗿𝗲 👀