The problem it really solved was your sysadmins were still operating by SSHing into the physical servers and running commands meticulously typed out in a releaae doc or stored on a local mediawiki instance, and acquiring new compute resources involved a battle with finance for the capex which would delay pretty much any project for weeks, while cloud vendors let engineers at many companies sidestep both processes.

Everything else was just reference material for how to sell it to your management.