I like the motor monitoring idea, however here is a different, unrealistic idea:
Heat the air above the column of water in the well. Put a thermistor in the bottom of the well and in the output water at the top of the well. Calculate the temperature difference and from that determine how long...