I tracked how various combs in the strain channel have changed in frequency over time using the autolines-complete.txt files for the daily FSCAN spectra (TSFT 1800s). I found that the 9.474, 9.475, and 9.480Hz combs have been changing in frequency over the course of O4, and have identified some key dates that warrant further investigation. I tracked the combs' frequency-roaming behavior by selecting a tooth which appeared often and was high-frequency. This was n=65 for the 9.474Hz comb, n=70 for the 9.475Hz comb, and n=64 for the 9.480Hz comb. The graphs of frequency over time I generated are available below. 

Observations:
The 9.474Hz comb's n=65 harmonic initially increases in frequency from 05-24 to 06-05, and then sharply drops from 06-22 to 06-24
The 9.475Hz comb's n=70 harmonic initially increases sharply in frequency from 05-26 to 06-05, and then sharply drops from 06-21 to 06-23
The 9.480Hz comb's n=64 harmonic initially increases sharply in frequency from 05-30 to 06-05, and then sharply drops from 06-22 to 06-23

So it seems that all three combs' frequencies begin increasing in very early O4 until 06-05 and then abruptly decrease from 06-21 to 06-24. It's likely that the combs were affected by changes made to the interferometer during these periods -- I believe that the week of 06-21 (Sunday) was when we set the power back down to 60Hz. 

I checked several other combs as well (n=11 of 4.98423, n=13 of 29.96518, n=161 of 1.6611, and n=3 of 11.9044) and none of those have observable changes in frequency (they may still be roaming inside their frequency bins) so this feature seems fairly unique to the 9.4Hz family.