If you set things up badly then you can certainly get paged when on AWS. If you set things up correctly (multiple AZs / regions and auto scaling etc.) then it's much less likely. There could of course be software issues but hardware is no longer your problem and you can let experts do it for you.

Yes, you could do this yourself but that would cost more and reduce the advantage. How many colo racks in how many data centres / countries do want to have to visit? Even with redundant hot spare fail-over you will need to fix the hardware at some point and will be running at risk until you do.

A hybrid approach can work, with colo for the base load and cloud for scaling spikes / redundancy.

> hardware is no longer your problem and you can let experts do it for you.

I've heard this before, but it's basically not true. You still have hardware problems, you're just paying to pretend they aren't there anymore. If you want to do that, it's OK, but we need to stop pretending that throwing money at a problem makes it go away.

> Yes, you could do this yourself but that would cost more and reduce the advantage.

Compared to AWS pricing? No thanks. I've yet to personally see a company save money by using AWS (ymmv of course).

> Even with redundant hot spare fail-over you will need to fix the hardware at some point and will be running at risk until you do.

Sure, but that's generally where you save your money. And if you do it right, it doesn't have to be at 4 am

At least with Google, you're paying for their transparent migration feature, which does make isolated hardware failures go away.

Most of these places have systems in place to make hardware issues unimpactful, which is why they make the big bucks. But you still have hard limits and issues that can and do arise because it's still reliant on hardware.