Unusual status of "predbat Demand (unhealthy)"

[Grey-Sheep]

Hi Each,
I have an unusual status on thePredbat Dashboard. It says "predbat Demand (unhealthy)"

There are no errors in the log, but a few warnings like:-

20544 | 2025-11-01 10:32:15.476854: Warn: Service call notify/home data {'message': 'Predbat status change to: Demand'} failed with response None

20248 | 2025-11-01 10:31:04.656487: Warn: Service call notify/home data {'message': 'Predbat status change to: Freeze exporting current SoC 79%'} failed with response None

19248 | 2025-11-01 10:10:50.516567: Warn: Service call notify/home data {'message': 'Predbat status change to: Exporting target 92%-77%'} failed with response None

15852 | 2025-11-01 09:00:38.735592: Warn: Service call notify/home data {'message': 'Predbat status change to: Freeze exporting current SoC 92%'} failed with response None

15361 | 2025-11-01 08:50:41.261336: Warn: Service call notify/home data {'message': 'Predbat status change to: Demand'} failed with response None

14388 | 2025-11-01 08:30:41.017700: Warn: Service call notify/home data {'message': 'Predbat status change to: Freeze exporting current SoC 92%'} failed with response None

5529 | 2025-11-01 05:30:53.285225: Warn: Service call notify/home data {'message': 'Predbat status change to: Demand'} failed with response None

4051 | 2025-11-01 05:00:52.615935: Warn: Service call notify/home data {'message': 'Predbat status change to: Charging target 85%-100%'} failed with response None

3451 | 2025-11-01 04:52:19.031356: Warn: Clip charge window 0 as it's already passed

It started (I think) after the last GIVTCP update, which also wiped out the GIVEVC integration (I had to delete the MQTT entries and let it rediscover)

Predbat is mostly fine but is making descisions as if it has lost the history, and is not good at adjusting in response to increased usage.

I have been looking for a few days and can;t see anything obvious.

Has any seen anything similar and could sujjest what I have done wring?

TIA

Jan

[gcoan]

those logged messages look like predbat is trying to send you a notification message when it has changed status, but the notification is failing.

Do you have a notify_devices entry in apps.yaml (you don't need to, but if you do, check this is correct).
Are you running the HA companion app which should automatically receive notifications (check the app permissions)?

[Grey-Sheep]

Thanks👍That was the underlying problem. Both an old phone left in the system, and the new one incorrectly listed!
The other anomalies appear to be connected to the GIVTCP update issues with the latest update (as reported in the Facebook group)