Energy values different between modbus and cloud dashboard

[davidwf15]

I have set up a raspberry pi with a modbus connection to the inverter for local monitoring, which seems to be working properly. However, some values are significantly different between the local reading and the cloud reading.
Trying to search online, I found this post which seems to describe the same issue but has no responses: viewtopic.php?p=13674

The images below show the values since the start of today and since the all time, for the cloud web page and from the inverter through modbus.

Screenshot_20260618_195133.png (26.16 KiB) Viewed 425 times

I have been trying to find which values in this document are the ones I want by matching with the cloud page (page 23, Table 3-6 Power and Energy data table): https://github.com/user-attachments/fil ... .04.00.pdf (I have struggled to find an original source of the document, but it seems to be correct).

The question is, why do some match exactly, some quite closely, and some not at all - especially on the all-time values?
Looking at multiple day's values, the export and pv production always seem to match exactly, the import is within a 0-500Wh, and the battery is always out by a few hundred. And this potentially compounds for the all-time values for the battery charge/discharge, but not the import?
The import value is usually consistently a few hundered Wh higher on the cloud page vs modbus throughout the day.

I think I have labelled the values correctly in my modbus output as the document's descriptions aren't very clear. I don't think "output", "enter" or "load" are relevant to what I want as they don't seem close to anything on the cloud page, and the home consumption seems to be calculated by adding all the supplies and subtracting the export and battery charge.

I don't really understand where the discrepancy can come from if I am getting values directly from the inverter via modbus, and the cloud gets values from the inverter via wifi. A small difference if the cloud has not updated yet, sure, but how can it get so far out over a long time?