This makes a lot of sense to me (as an Operations Engineer).

I could imagine the architecture team has low watermark triggers to rescale the architecture, kill and restore hosts, or other changes based on expected user load. When that load just… isn’t there, the automated tooling just loops the same actions causing site instability.

I’ve had similar issues before, so it seems like a feasible explanation