don't you think the capacitors are also integrated inside the ceramic pcb? I mean like they are eatched into the circuit board?
if they are underneath the attenuator, there is hope !
I've never taken a look into the shielded attenuators of these THS scopes myself, so I just was guessing. One can see a few capacitors soldered to the ceramic, especially the big one for AC coupling as I suppose. I've replaced the relays in my TDS580 attenuators (found some pictures here:
https://www.eevblog.com/forum/testgear/tektronix-tds520b-attenuator-swap/ ), here one clearly can see the large metallized areas - these are the capacitors for their attenuators, see the laser trimming. I don't know wether the THS ceramics are made the same way or they've soldered discrete capacitors. For sure, one can see laser trimmed resistors there (Edit: looked again at the high res picture, looks like a trimming scheme applied to whatever is hidden under the blue coating, can't tell if it's a capacitor or resistor).
also, can it be due to the relay as @Jim Narem suggests? Although i cannot figure out how the relay can cause imbalance capacitance
From my opinion / experience, a weak relay should cause DC errors. Within the TDS 5xx / 7xx attenuators, weak relays cause SPC errors, and/or complete loss of signal for some attenuation settings. Exercising the relays can cure that for a while.