To add to the above comments, yes, there is a "disaster." How exactly you do it, it depends on your BCP. For example, if you use a fault-tolerant data center, you will turn off the power.
You do not need to hard kill systems to do this. You can just pull out the net; really depending on what you are testing, which may be preferable.
However, what you want to do is periodic periodic, “disconnected”. My previous team was with a global financial services company. Our systems were 24x7x365 critical. But in fact, we needed to carry out whole data center failures - and they would go on for days. How often you do this depends on the results. No matter how well you simulate or emulate it, if you do not do it in production, it is mostly academic.
, , , " ". , - , , , , , . : . , , , - ; . BCP DR . , 3-4 . "" - , . . , - - .
. . , , .
, , . ( ), , , - , , . , , PTB, " " a) b) , .