@blyxyasAlejandra Gonzalez "If this limit was only reached in one machine, panicking only in that machine, should start a self-stabilizing mechanism to distribute the load of that machine into other machines"

Definitely, and I've also seen the question raised of why the update deployment system, when the deployment with the bad config file was pushed out, did not automatically roll back to the last known good configuration upon failure occurring. But neither of these, of course, are language-level criticisms.