As someone who occasionally has to deal with on-prem projects, I sometimes wish I could just use self-service and acquire VPSes as needed to solve the need for

As someone who occasionally has to deal with on-prem projects, I sometimes wish I could just use self-service and acquire VPSes as needed to solve the need for resources, as I can in my personal projects. Instead, I find myself being slowed down greatly: "Having a log shipping solution would be nice here, but I cannot set it up, because there's not enough RAM. Static code analysis? Not enough RAM. Another back end API instance for that new front end we're working as a replacement for the old one? Not enough RAM. Wanting to test how well scheduled processes run overnight with larger amounts of data? Too few CPU cores. Want more container images stored in your own Nexus instance? Not enough disk space. Want another database instance for CI or testing destructive migrations? Not enough of anything." Which is the problem that the cloud largely solves, at the expense of way much money spent than necessary. But hey, when you have to deal with red tape to get more resources on-prem, that's just a way of managing the overall process and manages risks that you might not care about in the moment when you "just want to get things done"; albeit that's not to share that you shouldn't care about them. So there are definitely problems with either approach (even cloud setups might have to go through someone and Jira/Redmine issues, as opposed to self-service), albeit at the end of the day most of those feel social/economical - after all, those large corps need their cloud service profit margins.