You have to assume that drives WILL fail from time to time... and SSD's aren't immune to failure either.
Back when I was doing this, we always kept spare drives in the cage with the servers. When needed, we could call the COLO staff and tell them to put a new drive into bay XXX and they would... then we remotely configure it and bring it online. Simple.
Exactly. If I have a drive failure on the other side of the ocean, we know immediately from automated notifications and simply have the duty engineer swap the drive with a new one which is inside the cage itself. It’s almost instantaneous.