Apologies for the necroposting.
I have some information that may well be of use regarding certain individual channels intermittently failing.
I've had the same thing, and earlier on in its life, it seemed to be corrected by cleaning the connectors with IPA and blowing some air into them to dislodge fluff etc.
Much more recently, I've found that this action didn't have much effect.
The nature of the fault is that over time various channels would come and go, almost as if there were some dry joints or one of those BGA ball problems. Some channels were more sticky than others.
Adjusting the threshold level on the LA has no effect, but if you supply an appropriate square wave into the dodgy channels, such as from the internal signal generator if it's a DS1000Z-S model, you usually see a glitch pulse on the LA at transitions, but no square wave.
Today I took mine apaaaaart, it'd come to the point where almost half of the channels had become unreliable.
There's a very short 68 pin IDC cable assembly with an IDC header soldered onto the main board. I discovered if I applied light pressure to the top of the header, it had an effect on the LA channels.
So I gently eased off the IDC ribbon header clip. After liberally applying IPA over the ribbon/blades, I individually firmly pushed each of the 68 ribbon wires onto each of the respective header blades with a 1.5mm flat bladed screwdriver I had to hand in an attempt to remake the cold weld.
I replaced the top clip of the header, and this seems to have resolved it, for now at least: it's been on for an hour with no ill effects, typically I'd have seen three or four channels randomly come and go in that time. It may be that I'll need to revisit if this problem recurs, perhaps to make the header clip maintain more pressure to the ribbon cable.
Edit: FWIW, over the years, I've pulled out and reinserted the RPL1116 cable while the scope's been on many times with no ill effects.
Edit2: The LA Probe Cal also passes after the above intervention.