There was a time when we attempted to conduct a "service cleanup" to eliminate those services that were constantly consuming computing and storage resources without being actually used.
However, the problem was that it was a crude era, with bloated organizations and chaotic management. We had no way of knowing whether these services were truly unused or who was responsible for them.
Ultimately, our solution was to randomly shut down a service and then wait to hear the roar of complaints about something was broken.
so, seems the admin heard the roar
that's chaotic engineering
