100% uptime is impossible, yes. But we're living in a world when your uptime is defined by how many nines you can place AFTER the decimal point. 99.9%, 99.999%, etc.
Having a fire in a data center, and having shit back up in a few hours; That's a company who has a disaster recovery plan in place. Losing a CPU and taking down everything on that server? For days? That's a company that's showing a lack of experience.
We're also living in a world where even a $5 a month plan gets backed up nightly with most companies that are at least decent. You pull the backups, load the VPSes (does a plural of that even exist yet?) onto other hardware, and swallow the performance hit until you can get the hardware back up and running.
I'm not criticizing Tilllmen, but his host is showing some serious "needs improvement" areas. But I don't trust that a bus couldn't just hit Tillmen one day; Bus drivers are maniacs.