Blog Posts

A couple of years ago, I worked in a SaaS company that suffered from probably all possible issues with software development. Code was so complex that adding simples changes could take months. All tasks and the scope of the project were defined by the project manager alone. Developers didn’t understand what problem they were solving. Without an understanding the customer’s expectations, many implemented functionalities were useless.

This post is a direct follow-up to Microservices test architecture where I introduced new kinds of tests to our example project. Wild Workouts uses Google Cloud Build as its CI/CD platform. It’s configured for continuous deployment, meaning changes land on production as soon as the pipeline passes. Note State of this article in 2026 This article is kept as an archive.

Thanks to tests and code review, you can make your project bug-free. Right? Well… actually, probably not. That would be too easy. 😉 These techniques lower the chance of bugs, but they can’t eliminate them entirely. Does that mean we need to live with the risk of bugs until the end of our lives? Over a year ago, I found a pretty interesting PR in the harbor project.

Do you know the rare feeling when you develop a new application from scratch and can cover all lines with proper tests? I said “rare” because most of the time, you work with software that has a long history, multiple contributors, and a less-than-obvious testing approach. Even if the code uses good patterns, the test suite doesn’t always follow. Some projects have no modern development environment set up, so there are only unit tests for things that are easy to test.

In the previous articles, we introduced techniques like DDD Lite, CQRS, and Clean (Hexagonal) Architecture. Even when used alone, they are beneficial. But they work best together. Like Power Rangers. Unfortunately, using them together in a real project is not easy. In this article, I will show you how to connect DDD Lite, CQRS, and Clean Architecture in the most pragmatic and efficient way.

You probably know at least one service that: has one big, unmaintainable model that is hard to understand and change, limits parallel work on new features, or can’t scale optimally. But bad things often come in threes. It’s not uncommon to see services with all these problems. What idea comes to mind first for solving these issues? Let’s split it into more microservices!

The authors of Accelerate dedicate an entire chapter to software architecture and how it affects development performance. One recurring theme is designing applications to be “loosely coupled”. The goal is for your architecture to support the ability of teams to get their work done—from design through to deployment—without requiring high-bandwidth communication between teams. Accelerate Note If you haven’t read Accelerate yet, I highly recommend it.

Did you ever hear about a project where changes were tested on customers you don’t like or countries that aren’t profitable? Or even worse: did you work on such a project? It’s not enough to say that it’s unfair and unprofessional. It’s also hard to develop anything new because you’re afraid to make any change in your codebase. In the 2019 HackerRank Developer Skills Report, Professional growth & learning was marked as the most important factor when looking for a new job.

I’ve seen a lot of complicated code in my life. Pretty often, the reason for that complexity was application logic coupled with database logic. Keeping the logic of your application together with your database logic makes your application much more complex, harder to test, and harder to maintain. There is already a proven and simple pattern that solves these issues. This pattern allows you to separate your application logic from database logic.

When I started working in Go, the community did not look favorably on techniques like DDD (Domain-Driven Design) and Clean Architecture. I heard multiple times: “Don’t do Java in Golang!”, “I’ve seen that in Java, please don’t!”. At the time, I already had almost 10 years of experience in PHP and Python. I’d seen too many bad things there. I remember all those “Eight-thousanders” (methods with 8k+ lines of code 😉) and applications that nobody wanted to maintain.