Try running this command to see if it is rebooting or bootstrapping a lot from the controller perspective :
show ap debug counters ap-name <apname>
We tested 6.3.1.0 with AP125's and didn't experienced reboots
What's your bootstrap threshold value , default is 8. You can try increasing it to 15
Those utilization numbers are pretty normal for the AP125 now , I see those numbers even on 6.2.1.3 with barely any devices associated to the AP
100.3.2-linseypool-ground LinseyPool-100 10.66.65.122 0 AP:HT:11/9/20.5 1 AP:HT:36/21/21 125
Peak CPU Util in the last one hour
----------------------------------
Timestamp CPU Util(%) Memory Util(%)
--------- ----------- --------------
2013-10-23 08:51:06 20 80
You should probably open a ticket with TAC too