Your expectation of the meter is the same as my expectation when it comes to the LED indication.
From what I have read, if the meter has power, the LED is on and the produces an 80mS off pulse for every Wh measured.
I had forgotten that the meter was upstream of the inverter isolator, therefore when I turned the inverter offf the meter has been turned off too, and so the LED turned off. (And the emonpi reporter measuring 3 pulses).
So this may not be a fair test, at the same time it is testing the emonpi and pulse counter aspect of the system to confirm the source of the problem.
I will try this overnight and see what the result is, if that is good then tomorrow night I will then try just removing the power cable from the bottom of the inverter as there will be no current through the meter, but it will still be powered.
Just one observation on timings:
It’s not perfectly every hour or half hour
3am this morning it was running about 26 mins past the hour
This evening 10pm it is running about 44 mins past the hour.
Will update again tomrrow!