The BS in yaw has an interesting pattern of wandering during a lock.  This typically starts out with a increase seen on the optical lever of about 0.5urad, then a decrease on the optical lever of about -1.5urad, and the an increase of about 1urad, which brings the signal back to about where it started, and the entire process takes about 10 hours, so it's slow.

This pattern repeats in multiple long locks, and there are number of lockes where the patern starts, but never finishes, because H1 loses lock.

The first two plots attached, each show 10 days (plot 1: Aug 8 to Aug 18, plot2: July 30 to Aug 8), and the pattern is visible in both plots. 

In the next 5 plots I've put t0 at each pattern, and saved the plot with a name that reflects if the pattern was complete or ended in Lock Loss.  In the most recent 10 days, there are 2 BS yaw wandering patterns that complete, and 3 that don't because H1 lost lock.

The last plot is of the wander on 10 Aug, 23:19:22UTC, which starts at around 9 Aug, 17:00:00UTC, and finishes around 10 Aug 05:30:00UTC, about 10 hours.