tl;dr They are often broken in pipelines and that might be because they are broken as a concept.

A typical issue in a microservices architecture is ensuring that the integration between services won’t break with the introduction of a new release. In my time as a consultant I have seen different organisations try to tackle this problem with varying degrees of success.

There is one particular approach that I have never seen fully succeed, although it is certainly popular. Organisations often try to have an “integration” test suite in a shared environment that spans several (if not all) services. It usually pokes and prods at functionality owned by multiple teams.

In this article I want to explore why I think these are a bad idea. I will also propose an alternative which is more in line with my understanding of the latest testing good practices.

First, let’s clarify the problem.