I checked the recent situation of the IPC glitch. 

• 12/13 n=2
• 12/14 n=0
• 12/15 n=5 (including longer IPC glitches)
• 12/16 n=0
• 12/17 n=0
• 12/18 n=1
• 12/19 n=1

I have attached the list summarizing the time the IPC glitch occurred in the JST timezone. In December, the glitch swarm seems to be concentrated between 22:30~23:45 JST somehow. Since the last glitch swarm finished on 12/8 at 23:45 JST, the work after or around this swarm will effectively reduce the IPC glitch. One possible work is [klog27901] to reduce the CPU load.