Daily utility meters did not properly reset last night

[Swallowtail (Simon)]

None of my daily utility meters (and there’s heaps of them) reset properly last night. These have not been touched in many months, but last night they all went to 0 at midnight, and then a minute or so later all reset back to the previous value.

This has never happened before.

Here’s an example, with 3 x daily cycle utility meters, all doing the same thing (you can only see the blue line down to zero and back, but they are all doing it):

This shows the last few days working OK and then last night:

I’ll see what they do tonight…

[thomasprior]

I think I’m seeing the same thing, albeit I had a period of “unknown” between midnight and ~8am when I started consuming gas again and the sensor jumped back to it’s previous value.

Container updated to 2024.9.2 last night.

[Swallowtail (Simon)]

So it’s not just me…

More granular view:

It’s affected utility meters setup both in YAML and also in GUI the same.

This is the definition for the one in this comment:

pv_energy_daily:
    source: sensor.pv_solar_energy
    name: "PV energy (daily)"
    unique_id: 2452716747910
    cycle: daily

I have hourly utility meters set off the same sources, and they appear fine… just all my dailies failed to rollover last night

[Swallowtail (Simon)]

Back to normal overnight:

[karwosts]

Suggest add a plot of the source sensor as well over the same timeframe, maybe something weird there.

[Swallowtail (Simon)]

I wondered that… But it was EVERY daily utility meter which did this (about 20), multiple sources.
I’ll check a couple of them but I’m not expecting every separate source to have triggered the same issue at the same time.

[Swallowtail (Simon)]

@Swallowtail

sensor.pv_solar_energy

Nothing unexpected in the source

[tom_l]

Can you zoom in on midnight for the source sensor history (23:55 to 0:05)?

[Swallowtail (Simon)]

Apologies, I missed the message - and the detailed history has now fallen out of the full window, however I did check at the time.

The source sensor (which in the example given was a solar array) had no entries at the rollover time, i.e. it was a straight line - as expected, as the solar array energy sensor did not accumulate any additional value from about 6pm until the next morning when the sun came up.

The issue has not recurred at this point.