> Or perhaps insisting on tiny VMs for your loads?
This happens way too often. Early-stage startups that build everything on the AWS free tier (t2.micro only!), and then when the time comes they scale everything horizontally
I'll repeat what I said above. It's for availability (aka I don't want my customers screaming at me if the machine goes down). And no , scaling out was not our first solution, scaling up was. I have considered going bare metal so many times, but the number of things we need to build/manage by ourselves to function is too much right now.
This happens way too often. Early-stage startups that build everything on the AWS free tier (t2.micro only!), and then when the time comes they scale everything horizontally