Today I made some temporary changes to the SQZ MANGER guardian, to turn off all the servos when the request is down. (Edited to add, now I've undone these changes but added a if False statement on line 270 in SQZ Manager's down state that can be toggled to put this back in).  This isn't usually what we want because we can leave the FSS and PMC locked. 

Derek pointed out in 81811 that some filter cavity length control channels are well correlated with the large glitches.  We know that blocking the squeezer doesn't stop the glitches, and without the squeezer locked to the IFO the filter cavity is locked on green and the noise of these channels seems too high to show the issue.

Request for operators:  If you see the glitches and low range happening again, please go out of observing.  First do a test of turning off and on the HAM1 feedforward (81775) it would be go to do 5 minutes off, 5 minutes on, 5 minutes off, then on again. Then open SQZ_MANager, and edit line 270 so that it says if True: , load sqz manager and request down, wait 5 minutes, request FREQUENCY DEPENDENT SQUEEZING for 5 minutes, DOWN for another 5, and back to Frequency dependent squeezing.  These things can be done at the same time that Robert walks around looking for noise if he's available when the glitches come back.