My server is still down - I wouldn't be surprised if mine will also show as a cold reboot, not least since it won't be the first time. Couple that with the issues my other server had at the beginning of the year (where the data store for a bunch of Afrihost VPSs was corrupted, and the best Afrihost could do was mount the old HDD into the VM so I could set up my VM from scratch again) and I have to wonder what is happening as far as redundancy/clustering is concerned.
Power issues should _not_ be an issue in a data centre - in these days of backup PSUs, UPSs, generators and dual power rails, power is just about the easiest thing to keep up and running.
I did learn from my experience at the beginning of the year, and now have my services backed up and fail-overed to various VPSs with several providers. Of all of them, Afrihost is the one provider that consistently has problems - and when they do, their recovery/explanations are extremely thin. When Isoho.st (a fantastic provider) found some data corruption in their store (before it became a problem, thanks to their monitoring) they were able to suspend their hosts and restore indices in about 20 minutes. All I had to do after that was update my system time, or wait for the ntp daemon to do it for me. This was accompanied with full details of what was happening and how, and it's the kind of disaster response that just makes me appreciate them all the more.
Tl;dr: Out of 6 different VPS hosts since the beginning of the year, Afrihost is the one provider that consistently drops the ball.