HandyHat/ha-hildebrandglow-dcc

data is an hour out

si458 opened this issue · 7 comments

si458 commented

Describe the bug
the data thats being pulled into HA is an hour out for some reason ?
im based in UK and its not day light savings time yet ?
the time is correct on my Pi too ?

To Reproduce
just setup the add-on and watch data come into HA

Expected behaviour
the data to be very similar BUT not an hour out?

Screenshots
Screenshot 2022-02-21 at 10 08 16
Screenshot_20220221-100758_Bright

Version
Home Assistant 2022.2.9
Hildebrand Glow (DCC) v0.5.2

Debug log

2022-02-21 09:04:05 DEBUG (SyncWorker_2) [custom_components.hildebrandglow_dcc.glow] get 2: (https://api.glowmarkt.com/api/v0-1/resource/b1f5af6e-3507-453c-ab8f-66ff9606b7de/readings?from=2022-01-01T00:00:00&to=2022-02-21T23:59:59&period=P1Y&offset=0&function=sum)
2022-02-21 09:04:05 DEBUG (SyncWorker_10) [custom_components.hildebrandglow_dcc.glow] get 2: (https://api.glowmarkt.com/api/v0-1/resource/5dbf33f3-1410-4334-a703-a3f9ad645a93/readings?from=2022-01-01T00:00:00&to=2022-02-21T23:59:59&period=P1Y&offset=0&function=sum)
2022-02-21 09:04:05 DEBUG (SyncWorker_11) [custom_components.hildebrandglow_dcc.glow] get 2: (https://api.glowmarkt.com/api/v0-1/resource/5dbf33f3-1410-4334-a703-a3f9ad645a93/tariff)
2022-02-21 09:04:06 DEBUG (SyncWorker_8) [custom_components.hildebrandglow_dcc.glow] get 2: (https://api.glowmarkt.com/api/v0-1/resource/b1f5af6e-3507-453c-ab8f-66ff9606b7de/readings?from=2022-02-21T00:00:00&to=2022-02-21T23:59:59&period=P1D&offset=0&function=sum)
2022-02-21 09:04:06 DEBUG (SyncWorker_1) [custom_components.hildebrandglow_dcc.glow] get 2: (https://api.glowmarkt.com/api/v0-1/resource/d6d9f3e9-2520-4150-85e4-9f4bc7d0ba9a/readings?from=2022-02-21T00:00:00&to=2022-02-21T23:59:59&period=P1D&offset=0&function=sum)
2022-02-21 09:04:06 DEBUG (SyncWorker_7) [custom_components.hildebrandglow_dcc.glow] get 2: (https://api.glowmarkt.com/api/v0-1/resource/c1c480a2-ab46-43eb-9d80-6135fa254763/readings?from=2022-02-21T00:00:00&to=2022-02-21T23:59:59&period=P1D&offset=0&function=sum)
2022-02-21 09:04:06 DEBUG (SyncWorker_4) [custom_components.hildebrandglow_dcc.glow] get 2: (https://api.glowmarkt.com/api/v0-1/resource/5dbf33f3-1410-4334-a703-a3f9ad645a93/readings?from=2022-02-21T00:00:00&to=2022-02-21T23:59:59&period=P1D&offset=0&function=sum)

Additional context
Add any other context about the problem here.

si458 commented

ive also just spotted my elec is also sometimes an hour out with data?
Screenshot_20220222-193923_Bright
Screenshot 2022-02-22 at 19 39 48

This seems to be the behaviour I would expect, and the best I believe we can achieve with this approach.
The DDC data is delayed by at least 30 minutes by DCC design. Thus any data from the Glow API is at least 30 minutes old.
When provided to HA it is provided as the “current” reading - there is no way (I know of) to say this is old data.
This HA will show it in the wrong time slot on the graph.

To avoid this, you need the Glow hardware, which will give current, not delayed data.

si458 commented

my Pi died of death, so i had to start a fresh, both my gas and elec data is still weirdly an hour out?

I'm seeing the same behaviour.

Was fine before daylight savings, now it's an hour out and shows my off peak costs as peak from 12:30-1:30am

si458 commented

@McMoobud im glad its not just me,
im still experencing it now even after the time switch, HA says 10-11 was 0.22 but that was actually 9-10 in Brights app
i also am missing any readings every day between 12am and 1am for both my gas and elec

As explained above, this is expected behaviour. The DDC data is delayed by at least 30 minutes by DCC design. Thus any data from the Glow API is at least 30 minutes old.
When provided to HA it is provided as the “current” reading - there is no way in HA to say this is old data - so it appear to be 1 hour delayed in the graph.
Thus HA will show it in the wrong time slot on the graph.

The missing data around midnight is also expected if using the daily sensor, as a result of the above issue. @HandyHat has reported in another thread he has spoken to Hildebrand and a solution is being worked on.

As of v1.0.0, the daily sensor no longer loses the last 30 minutes of the day. If you are still experiencing data being more than 30 minutes out, please say!