The ultimate test would be to pull a drive and find out

please make sure you have a back up before doing this, and also that you know what steps are required to return the array to an optimum (or synchronized) condition.
In theory, and in all my tests (yes I did pull a drive - more than once), the data remained available, an email was sent, and the appropriate drive LED turned amber. I will point out that if the drive were to fail whilst the unit were powered off, it is possible, depending on the firmware you are running, that you might not receive any any indication of drive failure, no mail, no amber LED, and only the blue LED not being lit to let you know that anything was amiss.
What do you do when/if it happens?
There's an FAQ on the support page with this information - but basically you back your data up, shut the system down, remove the failed drive and replace it with a new one of similar or greater capacity and then power the system up - when you log into the web admin page you will be prompted to format the new drive after which the system will rebuild the array.
It should be noted that some users have had problems with the wrong drive being formatted - especially when the "new" drive already contains data.
Last but not least - READ MY SIGNATURE