Comments on Amazon’s article
My colleague joked - the fiscal year has been ending, and OKRs should be achieved. That’s why the article was written.
The article tells how moving back to a monolith application, infrastructure costs are reduced. That’s true! A Microservice Architecture (MSA) application requires more spending than a Monolithic one. This cost is a tax for increased productivity, improved reliability, and scalability.
But, there is an extreme version of MSA is to use Lambda Function or similar when one function is called as a service. But this is not a microservice! I would call it a nanoservice. And this approach is used for small application (that requires multiple untied and autonomous functions), that are hard to scale.
And the article is about that case. The team has chosen the wrong architecture for their application, they met difficulties with scaling and highly extreme operational costs. They acknowledged the problem. Merged separate functions into one service, called a Monolith. But actually, it’s not a Monolithic app, it’s a good Microservice application :) And yes, it has a lower cost than nanoservice architecture )