That seems pretty similar to what I'm seeing. It did it three times in a row for me before I finally broke down and re-flashed the previous release.
Of those three times, the first one would still allow ssh sessions, but they were VERY slow to respond, it would take 20+ seconds to get the password prompt, and tasks I would perform were very slow. There was a process stuck at 100% CPU, but it wasn't doing anything, I couldn't kill it, and every ~30 seconds I was getting a message in the system log like I posted earlier. My system is running on a bench power supply, and normally the current draw is very dynamic, bouncing erratically between 200mA and 400mA at 10V as it's running and executing different tasks (as you'd expect). When this happened it was stuck at ~250mA solid, absolutely no fluctuation.
The other two times it happened I was locked out completely. I didn't have an active ssh session so I couldn't see what was going on, and any attempts to ssh into it just dropped. As before, the current draw was ~250mA solid, as if it was hung and not actually operational.
All three times it took around 60-90 minutes before this happened. A full power cycle gets it going again, until 1-2 hours later when it happens again. To reiterate, 20151204-36 does not have this problem.
This is the closest match I could find on Google:
It's a different kernel version, but the symptoms are the same. The cause in that case was a regression in the 4.1.9-1 kernel.