Thanks. I looked at one of the APs and sure enough, there is a keepalive-timeout.
Rebootstrap Information
-----------------------
Date Time Reason (Latest 10)
--------------------------------------
2018-11-01 06:58:01 Switching to LMS 10.1.27.5: KEEPALIVE-TIMEOUT. Last Ctrl message: HELLO len=1222 dest=10.1.27.5 tries=10 seq=0
Now I've got to figure out why as only about 20% of the APs have moved over and the other controller is still up.