Here's a few plots looking at the PSL glitches since the start of O4a. 

The first plot is similar to Oli's from 80520, blue shows the monitor at all times and orage shows only for times when the IFO was in a low noise state (600 or more) for the whole minute.  Since the blue trace includes transients from the locklosses there isn't much of an apparent trend looking at this data without masking out locklosses.  From the orange trace you can see that the maximum of the glitches that we have been riding out in low noise has gotten higher recently, after drifting somewhat over the run.

The next plot shows sort of a glitch rate, where I've counted minutes where the FSS fast mon channel goes above 0.6 V as minutes with glitches, and haven't made an attempt to try to count how many glitches are happening each minute. 

Aside: Thanks to Erik and Jonathan for helping me understand why Oli has been having a hard time getting data.  The reason was that gwpy no longer supports gap handling in the environment, as all the scripts in this duty cycle repo were doing.  Now we have to put the gap handling as an argument into gwpy, as shown on line 24  here