Affected instances have been restored. We are investigating the cause of the issue and will be making any necessary fixes to ensure this doesn't happen again.
