IMHO the odd spark/short shouldn't do any harm, but if you combine it with a bad bit of soldering on the production line or the last percentile component in the quality range it might be enough to send it over the edge.

The "exercise your ABS" plan is utter guff. The ABS does a full self test valve cycle at each ignition on (the clunk click as the lights change sequence on the older systems, more subtle on later ones) and partially energises the solenoids every few seconds to confirm they are electrically OK. Deliberately cycling the ABS on wet grass (a surface it is not designed for) is only good for sales of new body panels, levers, mirrors and medical insurance.

Changing the brake fluid as required is a good plan to get water out. Don't do it too often and leave the ignition off until you have a fully bled and working hydraulic set up, the fluid lubes the pump and valves.

The routing of the brake lights via the ABS ECU is highly suspect IMHO. The whole point of CAN is to utilise the info you have over the whole network, so telling the ABS to "wake up" via the brake light change of state is a good use that will improve its reactions. The fact is though that this signal can be picked up via the general network, the ABS can be a "listener" without having any direct input. Strikes me that the boys in Munich ran out of capacity on the instrument module and rather than expanding it decided to put the brake light reaction into the ABS ECU. Running the actual brake light wires a whopping 1 meter less saved them something too. I understand such logic on a bendy bus or a truck with potential road train uses, but on a 2m long bike it's ridiculous corner cutting.

Andy