Looking at the IMC transmitted light, I realized that the laser power occasionally dropped by 15% and came back to the nominal level in 10-ish seconds (see the attached screenshot). Since I knew this was usually associated with ISS, I took a look at ISS. Sure enough, the diffracted light was too low and it was about 5% or perhaps even less. According to the trend (see the second file attached), it seems that the diffraction power started decreasing after around 18th of December in the last year and recently dropped steeply. This too low diffraction was causing the frequent unlock of the ISS loop.

I adjusted the offset voltage to increase the diffraction power to 10% as was done by Doug (see alog 7810).

H1:PSL-ISS_REFSIGNAL = -1.93 ( which had been -2.00)

It seems this fixed the glitching issue as of now.