If it is sealed with a cap it fills up, and the water rises to the level of the greater outlet that goes to the pail. It's impossible to photograph this without damaging the maker so here is a diagram: Responses to concerns in remarks: [Edit 2] Leading theories Some really informing conversation in the answers and comments here.
Water gathers in the pan. It drains out in between cycles or, if there is enough humidity, it supports to the point that it drains out to the container. The pail exit hole is vertical, and it is generally bigger than the constant hole so it does not suffer the same problem.
in the pipe restricts outflow, normally not significantly however throughout periods of peak humidity the limitation suffices to trigger water to back up and overflow the pan into the pail. Perhaps the two above things integrate. Potentially gunk buildup near the continuous exit hole increases air inflow speed to intensify the problem.
Possibly flushing the exit hose periodically will help. Maybe by hand each year, or perhaps by linking the tube to the bucket instead of the pipe connection, as explained in among the answers. [Modify 3] An upgrade, one month later Edit 3 is 33 days after Edit 2. I have actually been viewing the container level sign daily.
The bucket remained empty throughout. Not a drop of water in the pail, ever ... until today. Between yesterday and today, the bucket filled and the device stopped. It was a bone dry day. It rained gently yesterday, nothing special. Temps remain in the 70s. I checked the pipe, and ran water through it from a garden faucet ...
The only remaining theory is unfavorable pressure. This is proven by my web tutorials on a similar issue with main A/Cs. https://paste2.org/6kpdfyN7 have created a makeshift "trap" for my system utilizing its cord keeper. It will be tough to state conclusively if this is working or not, considering that I can not correlate this with weather condition or anything else, and it's quite typical to go more than a month without the problem.