Thanks for the reply Scott. We've got replacement parts on the way it's just unsettling to think it's affecting six filers, and now anecdotally others too. Maybe a faulty batch? We tried upgrading firmware at which point the error changed to 'NVRAM battery status unreadable' instead. I'm guessing that it's a false positive alarm rather than faulty hardware but it's not a risk that's anyone would want to take. For the time being we've disabled the 'critical' threshold for NVRAM (which is 72 hours, warning at 80 hours) by setting 'option raid.timeout 0'. This at least prevents unmanaged controller failovers.
We'll continue to monitor the batteries once they're replaced and hopefully the issue won't recur.
Regards,
Ed.