GitOps at scale sucks. This equation roughly explains it P(gitopsfails)=1-p^n (humanitec.com)

You talk about scaling with GitOps is difficult. But isn't this an issue of scaling itself? Management overhead, communications channels (Conway's law). Scaling is always hard.

Well, yes, you might be right that the satisfaction of a team drops the longer you use a certain approach. But often, time affects quality. So I would argue that this is more of an organizational issue than a GitOps specific issue. Though, I can see the problems when scaling GitOps to 100+ applications or multiple teams working with this approach.

It's good to see some more critical analysis around GitOps from those that have used it at length, not only the initial hype benefits type of article.