BlueHost doesn't rate all that badly on the shared hosting chart (I am tracking this data - they are 6 out of 19 among some of largest providers). That sort of behavior sounds like crap, but it happens to most of them because of the nature of shared hosting. I am not sure that half a day to get a SQL load under control is normal though. I've had hosts just suspend my account right away when I go over or they kill the offending script.
That's what I figured they should do -- just kill the offending script automatically. They confirmed to me after several hours of outage that they had their best admin working on it. Not sure what there would be to work on that would take that long.
The shared hosting chart you mention sounds great! Is that available somewhere?
