I feel like I don't really understand the point of the article. Half of it explains the problem, and the solution in the other half is extremely vague.

It seems to imply that root cause analysis was previously only taking into consideration individual components and only machine-to-machine interactions, and a better approach is to also take into consideration how systems interact with each other, as well as human operators.

Maybe that's true for Google, but I don't think this is the "a-ha" moment the author thinks it is. If you are just looking at how two components can fail while interacting, then obviously you are leaving out a huge part of the equation. I don't see why we would need another aconym jargon for this.