Prior to the move to the cloud I had a half dozen 2U Dell servers running my SaaS in a data center. Great pricing, but I did (and still do) all IT-related functions. I had a 1/2 rack and bandwidth in a semi-secure colo. During that 5 years I was using them, I had one (1) minor network outage and a few unscheduled hardware outages to my own hardware, a couple of which required me to drive an hour to hard cycle one of the boxes (same box every time). I generally trust unto myself and that greatly pisses me off. I have several machines on a load balancer that are constantly synced and an 'emergency' backup in case the load balancer goes tits-up or the internal network goes down. Unfortunately, the hardware was getting old and I didn't want to replace it all for my own virtualized setup, so I looked around and went with Rackspace.
My regret is that Rackspace had some virtual switch issues that really affected performance and in several cases caused outage. It seems like they have cleared things up in their last update and I am cautiously optimistic, but remain vigilant. I run a PostgreSQL 9.1 cluster with no real master/master solution. I opted not do do something like DRBD for the database data due to speed issues and if I need to change to a new master all I have to do is kick off a script. That has never happened except while testing failover.
The bottom line is I believe in trusting unto myself. I'll happily use their hardware resources, but I don't like shared database environments so I have my own. I have several slaves in my DB cluster, have multiple forward-facing machines that are kept in sync and I keep a very close eye on things. Several times I have noticed larger than normal I/O issues that were being caused by someone else sharing the same hardware on one of my VMs. Sucky I/O really affects database performance a LOT and each time I notified Rackspace that they had a problem and they responded immediately, including after the 3rd time in the same morning, in which they disabled the VM of the idiot who was causing the issue until they could prove it was fixed. Their support is AWESOME and one of the main reasons I went with them.
My service is a small business solution geared towards the pet / dog daycare industry. There is a marketing component, but mainly scheduling and tracking pets. Short of the virtual switch issues Rackspace had with their Open Stack cloud, and the couple other instances, I have been more than pleased with their service.