techWithMatheus

Backend and cloud architecture.
Tradeoffs, failure modes, and real systems.

I work for AWS but opinions are my own.

[techWithMatheus @techwithmatheus]

Most microservices fail because of design decisions:
• Chatty sync calls
• Shared databases
• No ownership
• Treating it as just tech

#microservices #softwarearchitecture #ddd
New video ↓
https://t.co/roJT326x2p

[techWithMatheus @techwithmatheus]

Switching systems is the riskiest part of a migration.

One way to reduce that risk is using the Parallel Run pattern:

Run old + new side by side
Validate with shadow testing
Then switch with confidence

#softwarearchitecture #systemdesign
New video ↓
https://t.co/3GYvSkwKQT

[techWithMatheus @techwithmatheus]

Don’t migrate your monolith all at once.
That’s one of the fastest ways to introduce risk and cost.

The Strangler Fig Pattern gives you a safer path: replace functionality slowly instead of a full rewrite.

#microservices #systemdesign
New video ↓
https://t.co/J7BscCytg8

[techWithMatheus @techwithmatheus]

Building a modular monolith? Look into Spring Modulith.

For Java devs, this is your go-to. If you use .NET or something else, it's still a great reference for how to structure boundaries properly.

Super cool production-level stuff.

[techWithMatheus @techwithmatheus]

Are microservices always more expensive than monoliths? No.
Likely? Yes.

“Always” almost never fits architecture.

With the right design and scaling, costs can be similar or even lower.

A classic case of "it depends".

[techWithMatheus @techwithmatheus]

Will AI replace developers?

No. But... hard cold reality?

Yes, coding may become a commodity, just something AI does.
But that's a task, not the job.

That's why architecture matters more than ever. Learn how the pieces fit together and you're future proofing your career.

[techWithMatheus @techwithmatheus]

@pladynski I had a quick look at that and it's really cool! I take it that you work for Graftcode? Reminds me a bit of WCF and the SOA days when we generated proxy classes to make remote calls feel local but all 21st-century modernized without the overhead. great concept! 🕺

[techWithMatheus @techwithmatheus]

3 signs your modular monolith may be ready for microservices.

When scaling pressure, distributed data patterns, and team ownership friction appear, it may be time to evolve the architecture.

#microservices #softwarearchitecture

New video ↓
https://t.co/AX85SjM7EW

[techWithMatheus @techwithmatheus]

Modular monoliths are great… until scale hits.

Deployment coupling and wasteful scaling start appearing.
Sometimes extracting just the pressure points into microservices is enough.

#microservices #softwarearchitecture

New video ↓
https://t.co/KRv83qcouv

[techWithMatheus @techwithmatheus]

Ball of mud codebase?

You can evolve it into a modular monolith without rewriting everything.
Here are 4 repeatable steps to start structuring the system again.

#microservices #softwarearchitecture #dotnet

New video ↓
https://t.co/QsfTKpSrbo

[techWithMatheus @techwithmatheus]

Most monoliths start clean.
Then shortcuts accumulate, dependencies point the wrong way, and the architecture diagram stops matching the code.
That’s architectural entropy.
New video ↓
https://t.co/RUU1yrt4HR

#microservices #softwarearchitecture

[techWithMatheus @techwithmatheus]

A modular monolith only works if the boundaries are real.
Modules own their data, expose contracts, and keep changes isolated.
Otherwise it slowly becomes a ball of mud.
New video ↓
https://t.co/Y2CQcSrGIt

#softwarearchitecture #microservices

[techWithMatheus @techwithmatheus]

Monoliths are not the problem.
Ball of muds are.

Without boundaries, everything starts calling everything and hidden dependencies spread everywhere.

That’s how systems turn into a ball of mud.

#softwarearchitecture #microservices

New video ↓
https://t.co/q6V4oxy2s9

[techWithMatheus @techwithmatheus]

Computers were women.

Before electronic machines, “human computers” performed complex calculations. Most were women.

First algorithm: Ada Lovelace (1843)
First compiler: Grace Hopper (1952)

Happy International Women’s Day! 👩‍💻
#WomenInTech #TechHistory

https://t.co/OHJ0UkiLGc

[techWithMatheus @techwithmatheus]

Most teams jump straight to microservices.

But for most greenfield systems, start with a modular monolith instead.

You get the architecture benefits without distributed system complexity.

#microservices #softwarearchitecture

New video ↓
https://t.co/qwT4GtUher

[techWithMatheus @techwithmatheus]

When is a monolith the smarter choice?

Small team.
Unclear boundaries.
Low ops maturity.

#microservices #softwarearchitecture

New video ↓
https://t.co/26DvXtaYnl

[techWithMatheus @techwithmatheus]

Monoliths or microservices?
Wrong focus.

The real issue is whether the architecture was intentional.
Without discipline, either one will break.

New video ↓
https://t.co/m2GcgvSsh1

#microservices #softwarearchitecture

[techWithMatheus @techwithmatheus]

Microservices are independent…until they’re all behind the same API Gateway.
If it fails, everything fails.

Availability is part of your architecture, not an afterthought.
New video ↓

https://t.co/X9fiSB0jZS

#microservices #architecture #cloud

[techWithMatheus @techwithmatheus]

Scaling can fail because of your container registry.
If image pulls stall, services don’t start and recovery slows.
Container registry coupling explained.
New video ↓
https://t.co/inrPjiJdZ6

#microservices #devops

[techWithMatheus @techwithmatheus]

AWS Summit London 2025 is done! What a day!! 🎉

Was AWESOME doing a talk with Becky Weiss, VP/Distinguished Engineer, and John Oshodi,Senior SA, on how to use a data lakehouse to modernize your data landscape.

This is one of the biggest tech events are and it was electric! ⚡🥳