I was talking about a situation when prod is already down and something is better than nothing. For example, once we forgot to add timeouts to calls to a third-party service, everything worked great for a year, then suddenly the third-party server started hanging for no reason which made all our PHP processes hang, too, and those processes were holding DB connections so we went out of DB connections as well and the whole production went down. We quickly found the root cause and applied a hot fix (setting the timeout) by editing a PHP script on the prod server without waiting for the usual cycle "pull request -> code review -> deploy to stage -> testing -> release" which can take an hour.