This is a good question, tracking is done in the meter, afaik, as it has parameters to report kWh, but it should not halve the result in either case. kWh do not depend on whether or when the device is polled, except in how approximate the resulting kWh reading would be if the controller was tracking kWh based on wattage, and the metered watts would be identical whether the device is polled or not, and regardless of how often it is polled.
A kWh is a unit of energy, equal to the amount of energy used by sustaining 1000W over 1 hour.
A device operating at 1000 watts would need to be in operation for 1 hour to use 1 kWh.
A device operating at 10 watts would need to be in operation for 100 hours to use 1 kWh.
If you check power use now, it is telling you the wattage being used at the moment.
To provide an example: It looks like a reading was taken today of the wattage being used at 10:45 AM. At that time, 93 W was being used.
At 93 W it would take over 10 hours of sustained use to reach 1 kWh. Obviously power use now does not mean that 93 was sustained, that will shift throughout normal use, but if you were averaging somewhere around 125-150 W today, the kWh reading looks like what I would expect since about sunrise. It all depends on what is being used and for how long.
In other words, if you take a measurement of watts using “check power use now”, and it reports 150 watts, you leave all downstream power consumption the same for 1 hour, check via the power use now every once in a while and average 150 W, you will have used in that hour period .15 kWh.
Try pulling some current usage data every so often over an hour period and then check the updated kWh, does it match the math?