Possible methods for reducing data points for long term measurements?

I would like to use my DMM6500 to log measurements from a connected voltage output sensor over the course of months. The expected output of this sensor will cycling between 3v for a period, then dropping to 1v for about 300ms, and back to 3v, again and again. It is important that I verify and log that it always reaches this 3v, and always drops all the way to 1v.

I would like to measure and save fewer data points as much of the data will be redundant. The precision is only important to me within 100mV. And I only really need to save data points when the voltage changes more than 100mV, or passes outside of a window that is considered "passing".

I've tried a few combinations of trigger models, but I so far I'm only finding things that leave me with one trade off or another. Is there a simple way to only measure and save when the measured value changes by so much?