Fiddling around with several deployment and configuration management options I just came to the conclusion that there is too much code between me and the system

Fiddling around with several deployment and configuration management options I just came to the conclusion that there is too much code between me and the systems. I would like to have one bash script that I could start in such worst case scenarios that tests everything and tells me exactly what the problem is. Just imagine everything is down and you can not trust your dashboards. What would you put into such a script? What would you test for?