FWIW, the battery compartment is continuous to the switch (and driver?) board and the back of the emitter head board. That is pretty much the whole shebang exposed to water ...
Makes me wonder if it was the pressure-relief valve on the cap or the button.
Looking in mine, the button has a long shaft which presses on a surface mounted switch. Which means the switch is likely only sealed with a sliding dynamic o-ring. I would prefer a magnetically operated switch with a hard or static o-ring seal, which wouldn't actually be that hard to do, nor costly.