Engineering a good code base is very hard and is unique to every project and the mentality especially in the .net community of "loosely coupled code is good" usually ends up over engineering and adds more complexity than necessary.
I've seen far too many cases where added complexity has killed the project's time budget because things just take so much longer to do. You need to maintain a healthy balance and actually evaluate if writing this particular system within the project in a way where we can "swap it out" later on is actually a use case.
Every project is unique and "unit testing all the things" might not be the solution for your project. Where I work we shifted our mentality when we highlighted this problem to only unit test bugs and write less decoupled code and in bigger services that are more mission critical we do integration testing instead. This works very well for us but as highlighted earlier, every project is unique and you should experiment what works for your particular team because what works for others might not work for you.
I've seen far too many cases where added complexity has killed the project's time budget because things just take so much longer to do. You need to maintain a healthy balance and actually evaluate if writing this particular system within the project in a way where we can "swap it out" later on is actually a use case.
Every project is unique and "unit testing all the things" might not be the solution for your project. Where I work we shifted our mentality when we highlighted this problem to only unit test bugs and write less decoupled code and in bigger services that are more mission critical we do integration testing instead. This works very well for us but as highlighted earlier, every project is unique and you should experiment what works for your particular team because what works for others might not work for you.