I did state in the very first post that hourly billing is the only asset. Thats why its perfect for overflow. I can spin up a double extra large and use it for a few hours when I need it, and kill it afterwards. Thats it though.

I use r1soft for redundancy and the costs for that are nominal. That sort of handles any hardware issues for me. I can get back up and running in about 10 minutes from a catastrophic failure.

All the stuff you mention below really doesnt matter to me, and I cant imagine it matters that much to many people. All these tests you mention could easily be done off site, or mirrored to run locally.

Still not worth paying 3-4x for these random "abilitys"

For things that matter- Bandwidth, Processing Power, Ram- Dedicated servers have EC2 Beat, hands down. More power, Less cost.