Issue with Utility Meter - not matching underlying sensor

[lowebb (Brendan McAllister)]

Hi all, I am having an issue with my Energy Dashboard where a large spike is usually present in both consumption and solar generation.

Current setup is via GivTCP integration, and I’ve set up Utility Meters for import, export, battery, and solar values, each with ‘Periodically resetting’ enabled.

However, if I compare the underlying sensor history to the utility meter sensor, the utility sensor does not always reliably reset back to zero

So it seems the underlying sensor is working correctly, but the utility meter isn’t, and I can’t figure out why?

[IOT7712]

Can you post your yaml code [inside </> formatting for correct spacing and readability]

[lowebb (Brendan McAllister)]

Assuming you mean the yaml for the utility meter? Thanks!

{
  "created_at": "2025-11-15T16:50:30.619820+00:00",
  "data": {},
  "disabled_by": null,
  "discovery_keys": {},
  "domain": "utility_meter",
  "entry_id": "xxxxxxxxxxxxxxxxxxxxxxxx",
  "minor_version": 2,
  "modified_at": "2025-11-15T16:50:30.619848+00:00",
  "options": {
    "always_available": false,
    "cycle": "daily",
    "delta_values": false,
    "name": "Grid Energy Export Today",
    "net_consumption": false,
    "offset": 0,
    "periodically_resetting": true,
    "source": "sensor.givtcp_xxxxxxxxx_export_energy_today_kwh",
    "tariffs": []
  },
  "pref_disable_new_entities": false,
  "pref_disable_polling": false,
  "source": "user",
  "subentries": [],
  "title": "Grid Energy Export Today",
  "unique_id": null,
  "version": 2
}

[Fanful]

It’s the source sensor perhaps jumping at midnight between 0 and the previous value and them back? That would explain what we see in graphs. I’d check with the graph zoomed to a period of a minute around the midnight when the discrepancy happened.

In general that would be my expectation, that the source sensor does some weird jumping.

But also, you should be able to use the source sensor directly in the dashboard, no need for the utility meter for this, normally. Why did you create the utility sensor? But of course if the source sensor is the issue, using directly in the energy dashboard will get you the same result as now.

[lowebb (Brendan McAllister)]

Yes you are spot on! The underlying sensor reverts to zero, then spikes, preventing the UtilityMeter from resetting correctly. Great catch, thanks!

I guess I’ll have to poke that integration and hardware to see if anyone knows why that is happening