Having spent the last two days monitoring the server we are now
confident that the early morning sickness has been resolved. We
updated the Linux kernel/libraries last night and ran the evening
backups manually. The backups completed without issue and the server
did not re-mount the hard drive as "read-only". This issue is now
closed.

 Please get in touch if you have any questions.

We replaced the PSU and SATA cables on our main server last night and
as a result web services were down for a very brief period between
10:00 and 10:20. Upon further investigation we now believe these not
to be the cause of the problem.
 
We have narrowed the problem down to either load related issue, or
something that occurs during early morning backups. The backups that
we are running tonight are in debug mode, so we should have a large
amount of info should the problem happen again before the backups
complete.
 
If you have any troubles with your sites today then please get in
touch. It is likely that this problem will be solved during the course
of today and while we locate the cause of the error, the server
continues to operate as expected.

We will be taking our main server off-line within the next hour to
replace both the power supply and the hard drive cables. This
emergency maintenance is being carried out in response to an issue
that occurred on Monday morning and again today. We expect the
downtime this evening to be brief. If you have any questions please
contact us via email.