Microservices are hard Build boring stack first
Given the hype for microservices, it's ok to become crazy about it and copy questions of why/what of likes of Netflix/Uber, etc. But at the same time, you need to understand A Netflix/Uber style microservices aren't required by (most?) org. They build it because they understand the domain they’re working in and knows their answers to how. Your approach to building one, without understanding your domain, or if you’re still trying to figure it out, microservices could do more harm than good because
1. Building reliable and testable microservices is a lot harder than you think. Distributed Transactions, synchronized clock, tracing, observability, debugging, etc are never easy. 2. Effectively testing microservices requires a ton of tooling and foresight.
So build boring stack first, Understand your domain and tech-sack. Monoliths are a good starting point to do so. Scale with monoliths first, figure out what is the bottleneck and optimize (A solid monolith can easily handle 10k rps nowadays). You don't become Google/Netflix/Uber, etc scale from day one. By the time you get there, you will have enough knowledge to build one.
No comments yet.