Practical .NET and Software Architecture Tips | Microsoft MVP

Hi there! 👋

I'm a senior software engineer, primarily working in the .NET ecosystem.

Things I talk about on LinkedIn:
- C# and .NET ❤
- Software engineering and architecture 
- Distributed systems and how to build them
- Databases and optimization techniques
- Career and personal growth

One of my goals (i̶s̶) was to become a Microsoft MVP in Developer Technologies.
I became an MVP in March of 2023.

If this sounds interesting, let's connect!

Milan Jovanović

I upgraded my Clean Architecture template to .NET 10.

And it took only a few minutes. Here's how 👇

My .NET version is defined in one place, a Build props file. I updated to net10 and that took care of all the projects in the solution.

The next part was updating the NuGet packages.

This was equally simple with central package management. I updated the version in the Packages props file, and it was picked up by all the projects in the solution.

Here's what's inside:
- Simple domain model with Domain Events pattern
- CQRS (no MediatR) + clean use cases implementation
- Validation and logging cross-cutting concerns
- Authentication with JWT and user context
- Boilerplate for RBAC authorization
- Docker Compose orchestration
- Minimal APIs, health checks

You can get the template here: https://lnkd.in/eihaFihr

I recently found out my template is used by a local company, doing $30M+ yearly revenue.

Did you check out the new IExceptionHandler feature?

You have to handle exceptions somehow.

IExceptionHandler implements the "try-" pattern.

You can decide which exceptions to handle.

To tell the middleware you handled the exception, you return true.

So, what makes this so interesting?

You can also chain multiple exception handlers that will run sequentially.

The first one to handle the exception successfully breaks the chain.

Here's how to use this in .NET 10: https://lnkd.in/e9k5yeJ3

It's a way to compose multiple exception handlers.

How do you approach global error handling?

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/e3q7Nu4s

LinkedIn

JSESSIONID

li_at

linkedin.com

Talks about #csharp, #dotnet, #databases, #softwareengineering, and #softwarearchitecture

.NET 10 is officially released! 🎉

Here are the top updates 👇

First, .NET 10 is an LTS (long-term support) release. So we get support for the next 3 years.

🔥 Performance improvements

Under the hood, .NET 10 delivers major performance boosts. The JIT compiler has been tuned with new optimizations such as de‑abstraction, smarter inlining heuristics and constant‑folding, allowing the runtime to do more work at compile‑time rather than at run‑time. There are many other optimizations across collections and LINQ, which translate to tangible speedups across everyday workloads.

🚀 Aspire goes polyglot

Formerly .NET Aspire, Aspire is now a polyglot, developer‑first platform for orchestrating front ends, APIs, containers and databases. Version 9.5 introduces a preview of a single‑file AppHost, letting you define and run your entire distributed application in a single file. The release also debuts a Generative AI visualizer to inspect and debug prompts and responses, plus expanded AI support with typed clients for GitHub Models and Azure AI Foundry and first‑class OpenAI integration.

🔗 EF Core’s new LeftJoin/RightJoin operators

LINQ’s lack of a built‑in outer join has long been a pain point. EF Core 10 fixes this by adding LeftJoin and RightJoin operators, so the code now reads exactly like the SQL you intend. For example, you can join products to reviews using LeftJoin() and EF will generate the correct LEFT JOIN statement under the hood. A corresponding RightJoin() keeps all rows from the right side while bringing in matches from the left. These new operators eliminate the verbose GroupJoin + DefaultIfEmpty pattern and make LINQ queries easier to read and maintain.

What's your favorite feature coming in .NET 10?

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/e3q7Nu4s

What's new in .NET 10?

I failed an interview because I didn't know how locks work in C#.

Read this to make sure it never happens to you.

I've told this story a couple of times before, but it's worth repeating.

I was in an interview some 7-8 years ago, doing a live coding exercise.

I was integrating with a currency conversion API and needed to add caching.

The problem was how this would scale with multiple concurrent requests.

A classic cache stampede problem.

I wanted to prevent this with locking.

Locking allows us to control how many threads can access some piece of code.

Note that I rarely worked with low-level code.

But the company I was interviewing for does.

Of course, they'd have this in the interview process.

Here's the situation:

- I knew I needed a locking mechanism
- But the method was async (oof!)
- So I tried to solve this using the `lock` statement

And that's where it all went to shit...

The `lock` statement doesn't work with async-await.

And it makes sense.

You can't guarantee which thread will complete the method.

So what's the solution?

You need to use an asynchronous synchronization primitive.

Luckily, C # has many to choose from.

Semaphore, SemaphoreSlim, Mutex, Monitor, etc.

Two important remarks:

1) Use a timeout when waiting on the semaphore
2) Always release the lock with a try-finally

Here's a tough question for you.

How would you implement a lock in a database?

If you want to find the answer, go here: https://lnkd.in/eVTMfu-y

Would you have nailed this interview?

---
Sign up for the .NET Weekly with 75K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/ejtkZwup

You can find security risks in your code at build time.

Here's a simple way you can implement this in .NET in a few minutes.

Imagine being able to find security risks when you build your code.

Static code analysis lets you do this.

Here's an example build error that warns me my password hashing code is insecure. 👇

Static code analysis can help you write secure and consistent code. And you can also enforce this at build, even in CI pipelines.

If you want to see how to set this up from scratch, check out this article: https://lnkd.in/dyYFpBYZ

Note that this doesn't detect every possible issue.

If you want to take this further, you can look at more powerful solutions like SonarQube.

---
Sign up for the .NET Weekly with 75K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/ejtkZwup

I love working with the Service Discovery library in .NET.

It's a lightweight configuration-based approach.

This actually came out of the effort done in Aspire.

If you're new to this concept, here's a quick breakdown.

Service discovery allows developers to reference external services using logical names instead of physical IP addresses and ports.

There's a centralized store for services to register themselves.

Clients can query the service registry to find the service's physical address.

The service discovery library for .NET is simpler, though.

It still lets you use logical names to refer to external services.

However, the discovery side is "lacking".

You have to specify the physical addresses in the application settings.

Here's a step-by-step tutorial: https://lnkd.in/eX4Wy3jD

You can configure the service URLs during deployment in your CI/CD pipelines.

Or even use a central configuration store (Azure App Configuration) for service discovery URLs.

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/eihaFihr

Most REST APIs fail not because of bad code, but because of bad design.

After building and reviewing dozens of APIs over the years, I’ve learned one thing: Consistency and pragmatism matter more than “perfect” architecture.

In my latest video, I share 8 pragmatic REST API design tips that I use in production. From consistent naming and correct status codes, to pagination, error responses, and hypermedia links, all shown in a real .NET example app.

Check it out: https://lnkd.in/enzqCKDs

If you’re designing APIs for real-world systems (not toy projects), this one’s for you.

No alternative text description for this image

If you're still writing five separate interfaces and classes for every single command in .NET, you need to see this.

For years, many of us in the .NET community have relied on MediatR and MassTransit for command handling and messaging. But the boilerplate is real.

I finally took the plunge and explored Wolverine, a library that claims to be the "next generation" of .NET messaging by focusing on convention over configuration.

In my new video, I share my raw, first-take as a long-time user of other messaging tools: https://lnkd.in/eDm9Aa3p

What's the fastest way to do a SQL bulk insert?

I tested out all the popular approaches...

SqlBulkCopy is your winner. It's the fastest way to insert a large amount of data into SQL Server. But it does require writing more code than the other approaches.

Postgres has the COPY command, which works similarly.

Here are 5 more solutions I compared: https://lnkd.in/eearSh_z

Which one would you use?

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/e3q7Nu4s

The fastest way to do Bulk Inserts

Need to generate a PDF document?

I love the simplicity of using HTML templates to generate PDF documents.

I create a Razor view as the template and render it to HTML with my data.

You can also use Handlebars. It doesn't matter as long as you create an HTML document.

Then you need a library that can convert the HTML into a PDF.

I used IronPDF in this example. Free options include Puppeteer or Playwright.

Here's a simple way to get started: https://lnkd.in/ewai8kjk

This could be all you need for generating your PDF documents.

What is your approach to creating PDFs?

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/e3q7Nu4s

PDF Reporting in .NET

Stop storing secrets in appsettings.json. Seriously.

That file was never meant to hold production credentials, yet I see it in real projects all the time.

I just published a new walkthrough where I show how to secure your .NET apps with Azure Key Vault:

- Creating your first Key Vault
- The exact RBAC roles you actually need
- Storing secrets with versioning
- Authenticating with DefaultAzureCredential
- Pulling secrets directly into ASPNET Core configuration
- Loading connection strings + options without touching appsettings

It’s a clean setup that keeps your sensitive values out of the repo and follows Azure’s best practices.

Learn more here: https://lnkd.in/eU4mUKnY

If you want to upgrade how you handle secrets in .NET, this one will help you get it right from the start.

I built a stock price dashboard using Cursor + .NET. 🔥

The real-time component uses SignalR.

The backend gets the initial stock price from a free stock market API.

Then, I have a background job that generates a price update and streams it to clients using SignalR.

The UI is pure HTML + JavaScript + Tailwind CSS for stylization.

And I vibe coded all this with AI in a couple of minutes.

It took me about ~15 seconds to add support for dark mode.

Did you have a chance to try Cursor?

It's my go-to AI editor these days.

Developer + AI is the future.

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/ew5Adz89

Are you accidentally crippling your EF Core queries?

Before you start bashing on EF, this is an SQL problem, not an EF problem.

But it can be problematic to find and fix with EF.

Understand the Cartesian Explosion problem and how it occurs in your queries.

I'll give you a practical solution using EF Core's query-splitting feature.

Here's the full story: https://lnkd.in/ezv5B48v

In future versions of EF, the split queries will be batched, which should significantly improve performance. 

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/eihaFihr

EF Query splitting

I rewrote a 40-year-old APL system to a modern tech stack.

It powers a $10M manufacturing business with 460+ database tables.

Here are 5 lessons I learned in 4 years:

1. Start with core business logic first

2. Modular architecture is worth the investment

3. Data sync is harder than you think

4. Balance technical needs with business pressure

5. Document institutional knowledge before it's gone

Want the full story of our legacy migration journey?

Check out my blog post: https://lnkd.in/eacxqj7z

This project taught me so many things about enterprise architecture.

I think you'll learn a few things too.

---
Sign up for the .NET Weekly with 75K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/ejtkZwup

Looking for a simple message bus in .NET?

I recommend taking a look at System Threading Channels.

You can implement a simple in-memory message bus using Channels in .NET.

Channels implement the producer-consumer pattern:

- Producers asynchronously produce data
- Consumers asynchronously consume that data

A channel has a Writer and Reader, that you can use to publish and consume messages. respectively.

Here's how you can use this to implement a simple message bus: https://lnkd.in/ekfJZd5n

Remember that this is still an in-memory bus, so it's not reliable.

You can implement resiliency around this, but it only goes so far.

---
Sign up for the .NET Weekly with 75K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/ejtkZwup

Most teams get microservices communication wrong.

They obsess over APIs and forget the real problem: coupling and reliability.

In my latest video, I break down how microservices should actually talk to each other. You’ll see where synchronous (HTTP, gRPC) calls make sense, and where you should switch to asynchronous messaging with queues and topics.

I also cover real .NET examples with Azure Service Bus, Redis, and distributed tracing, so you can see how to design communication that’s fast, reliable, and easy to reason about.

Check it out here: https://lnkd.in/eGB3WUqW

If you’re building or refactoring a microservices system, this will save you a lot of pain later.

Implementing resilience in .NET? You better...

Polly is probably your first choice.

And Polly is much faster in V8+.

But why should you care about making your application resilient?

Let's say your services communicate by sending HTTP requests.

HTTP requests can fail because of network or server issues.

These failures can disrupt service availability, especially as dependencies increase. This increases the risk of cascading failures.

Luckily, there's a simple way to solve this in your .NET applications.

Here's how you can make your apps resilient: https://lnkd.in/e39__DXH

Do you use Polly in your applications?

There's also an official resilience library in .NET.

But it's a wrapper around Polly that adds OpenTelemetry on top.

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/eihaFihr

Was API Versioning a mistake?

I built and maintained a large public API (400+ endpoints).

The API has dozens of integrations, serving mainly mobile applications.

When your API is serving so many clients, breaking changes are expensive.

So, everything I implemented on the public API had to be planned.

If I wanted to introduce breaking changes, I had to version the API.

This wasn't a standard I decided on, but I had to live with the consequences.

API versioning allows your API to evolve independently from the clients using it - or so they say.

What we actually need is change management for APIs.

1. Don't remove anything
2. Don't change business rules
3. Don't make optional things required
4. New things must be optional

I explained this (and many other things) in-depth in Pragmatic REST APIs: https://lnkd.in/eXP53TP6

What do you think about API versioning?

P.S. I've got a Black Friday offer for my Pragmatic REST APIs course that expires soon. Don't miss it.

---
Sign up for the .NET Weekly with 75K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/ejtkZwup

Do you know all 3 service lifetimes in .NET?

ASPNET Core has three service lifetimes:
- Transient
- Singleton
- Scoped

Here's how they work. 👇

Transient services are created each time they're requested from the service container.

Scoped services are created once within the scope's lifetime. For ASP .NET Core applications, a new scope is created for each request. This is how you can resolve scoped services within a given request. You can also create a custom scope within a method, for example.

ASPNET Core applications also have a root service provider to resolve singleton services.

What if you want to resolve a scoped service inside a singleton service?

Here's what you will need to do: https://lnkd.in/eg_EQHb6

The solution isn't too complicated.

Save this for when you need it.

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/eihaFihr

This skill will make you valuable in any team.

Diagnosing and solving performance bottlenecks in your system.

I always strived to make my application as fast as possible.

So, performance problems are some of my favorite things to work on.

In microservices systems, this problem becomes even more prominent.

Why is this the case?

Finding the performance bottleneck is much harder in distributed systems.

Luckily, you can track the flow of requests through your entire system with tracing and OpenTelemetry instrumentation.

I wrote a step-by-step guide for configuring distributed tracing, and I'm confident you can apply this in your applications.

Here's how to get started with tracing in .NET: https://lnkd.in/eZ76HXr6

Are you integrating OpenTelemetry into your applications?

---
Sign up for the .NET Weekly with 74K+ other engineers, and get a free Clean Architecture template: https://lnkd.in/e3q7Nu4s

ANDMORE

Software Engineering Manager

Levatas

Development Team Lead

Senior Software Engineer

EndToEnd Management

Senior Software Engineering Consultant

Laura Software

Senior Software Developer

Independent 

Independent Software Developer

Software Engineer Manager at Juniper By IMC

Software Engineer Manager with over 15 years of experience in software development and three years of experience in management. Proficient in various programming languages, including [C#, Node, Javascript, and Python]. 

Experienced in leading cross-functional teams to deliver complex software projects on time and within budget. 

Strong background in agile development methodologies and project management.

Extensive experience in the entire life cycle of the software design process, including requirements definition, prototyping, design, implementation, testing, and maintenance.

Raul Junco

You don’t need microservices. You need paying customers.

Scaling problems are good problems because they mean success. Too many teams over-engineer before they have real demand.

Before breaking things into microservices, ask:

• Do we have performance bottlenecks, or just assumptions?
• Is our monolith slowing us down, or is it poor code structure?
• Is our database the real bottleneck, or are we misusing indexes and queries?
• Are we solving business problems, or just adding complexity?
• Do we have the team to run microservices?

Microservices won’t fix a bad product. 

A well-designed monolith scales further than you think. 

Get users first. Optimize when it hurts.

Anyone can return 200.
Real engineers return 202 and still deliver.

Most teams think “faster” means “simpler.”
They block until everything’s done.
Then they wonder why latency spikes and the system melts under load.

A. Make it fully synchronous. “Just wait till it’s all complete.”
Safe on paper. Pain in production.

C. Write to the database and fire off background jobs.
Feels fast. But when those jobs fail, no one knows until it’s too late.

D. Run a saga in the same request.
Total control — until one step hangs and the whole thing dies.

The real move: 202 Accepted + durable queue + status resource.

Why it works:

1. Clients get instant responses. No waiting.
2. Queues absorb spikes, retries, and downstream chaos.
3. You gain visibility through status endpoints and webhooks.
4. Systems stay decoupled and fault-tolerant.
5. Latency becomes predictable; you own the timing.

Trade-offs worth paying:

- Eventual consistency. Things settle, not sync instantly.
- Job tracking. You need a clear state machine.
- Idempotency. Every retry must be safe.
- DLQs and metrics. Otherwise you’re blind when it breaks.
- Ordering. Version every update or lose sequence.

200 shines in happy paths.
202 shines when the world’s on fire.

So here’s the question:
When do you let your API walk away instead of waiting?

Every system has a hot path.

And it’s the first thing you need to understand before you scale anything.

Here's how to scale smart:

1. Start with the access pattern

If reads massively outnumber writes, don’t scale them together.

Example:

E-commerce app

- 99% traffic → product search & browsing
- 1% traffic → product updates

Split them.
Let read-heavy services scale independently.
Keep write flows lean and safe.


2. Optimize storage for the real bottleneck

One primary DB for writes.
Multiple replicas for reads.

- Writes stay consistent
- Reads stay fast
- Spikes don’t affect the source of truth

This is simple, and it works.

3. Add a cache where it hurts

Not everywhere. Not blindly.

Cache:

- Product pages
- Search results
- “Top N” lists
- Anything that gets hammered repeatedly

Memory beats disk.
Microseconds beat milliseconds.

They don’t teach this in bootcamps.
Because the real rule isn’t “add Redis” or “use replicas.”

The real rule is:
You scale what’s hot; not the whole system.

taplio

The Battle

Milan Jovanović

Raul Junco

Followers

Posts

Engagements

Likes

Performance ScoreMilan Jovanović's LinkedIn Performance

Performance ScoreRaul Junco's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Milan Jovanović

Raul Junco

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Milan Jovanović's LinkedIn Performance

Performance ScoreRaul Junco's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreMilan Jovanović's LinkedIn Performance