Figure 1. Yesterday the power on B1p in CARM NULL 1F became consistently higher, increasing from typical 300-400mW, to 500mW or more. At the same time the 56MHz sideband balance on B1p has changed, instead of having the USB with about double the power of the LSB in CARM NULL 1F, the USB power has started to drop off very quickly after reaching CARM NULL resulting in a misbalance in the opposite direction, with LSB having double the power of the USB. 

Adding a very large offset of 75 on MICH SET could restore the sideband balance (around 17:30 UTC on Oct 20), and decrease the power on B1p to 400mW, then the addition of the DARM offset has increase it to ~450mW, which are still both higher than normal. 

Figure 2 the change in behaviour seems to have started between the lock at 5:00 UTC and the one at 8:45 UTC. But there has also been a normal looking lock at 12:30 UTC. I haven't been able to find a reason for that change. The first occurance of it is before the inspections of the cabling around the CO2 lasers https://logbook.virgo-gw.eu/virgo/?r=67991

Figure 3. Another witness of the issue is SRCL INPUT that has become systematically lower by a factor few. That behaviour looks to have started at ~6:20 UTC, at the same time as the sideband balancing changed during a LN3 lock.

Figure 4. At 6:20 UTC there was the WI DAS that dropped off in power for 15 seconds. A short power cut like that did not have a large and long lasting consequence during the weekend, so it cannot by the reason by itself. Something else must have changed at the same time.