The server went down this morning due to an audible alarm coming from the RAID controller. I spent the morning investigating what had gone wrong and getting the failed drive replaced. It turned out to be the same drive that had failed previously. As before, it actually came back up and was successfully rebuilt by the RAID controller, which I find a bit odd. The same hard drive works perfectly for a few weeks, then drops out, but rebuilds upon simply reseating it. To me, that sounds like some sort of vibration/connector issue, but one of the guys at the datacenter said it could also be a symptom of a drive that is gradually failing. So, I decided to just replace it with my new spare, and send it back to WD for replacement under warranty (I think we're still in that window, these are server-class drives with 5 year warranties). I'll see if the new drive performs without problems, if so then it probably was the drive. If the same thing happens again then I'll have to look at the connectors and/or drive bay hardware.
Anyway, the server is back up again now. Apologies for the downtime, at least this time I was able to go in there quickly and deal with it myself. Hopefully it'll stay up.
Thanks for your patience,
Neil