This is a continuation of previous work (klog#19859).
I performed the Hveto again, by considering several comments from Ushiba-san and Akutsh-san.
Here are the updates.
* I added the channels related Green laser control and PR3 oplev. For the used channel list, please check here. The unsafe-channels are channels that are considered to be correlated and have been removed from this analysis.
* From the Hveto result, the channel (K1:IMC-SERVO_MIXER_DAQ_OUT_DQ) was the round winner which is the most coincident channel with the lock loss of GreenX.
* As a quick check, glitches on K1:IMC-SERVO_MIXER_DAQ_OUT_DQ happened earlier than glitch on GreenX (K1:TMS-X_GR_PDA1_OUT_DQ). These were done by checking the glitch time obtained by Omicron. But, this time seems to have an uncertainty. We should confirm this by the other reliable way.