I observed this this morning, too. I also have a 2m packet observation that looks similar:
(this could be another satellite, maybe AISAT, but I won’t know until later today when I have a chance to look at keps and try and decode the audio with direwolf).
In any case, I started to notice weird stuff with the PSK downlink a few months ago and posted to the AMSAT mailing list about it. A rep from Brno University responded saying that this strange behavior is probably related to low battery voltage due to the batteries deteriorating and the voltage dropping when in eclipse.
my guess is that battery issues are indeed the cause of such oscillations.
For them to be Doppler shifts due to spinning, the satellite would have to spin really really fast.
With some quick calculations, in order to observe a 5khz doppler shift due to spinning you will need NO-84 to spin at around 200k RPM