More specifically, a differential backup of the database and file structure occur every evening and the full backup occurs every Sunday. We have a parallel staging environment to ensure that all data is being backed up properly.
We are fully redundant in our collocation facility. If the server fails, there is another server ready to come online seamlessly. If the CPU, hardware and RAM fail, again there is minimal to no downtime for our users.
For specific liability questions, please see the Terms of Service here.
You may also wish to add your voice to a related Idea: Read-only access during system updates, as this suggestion is currently under consideration by our team.