An option to not show PV meter error for "sleeping" at night inverters

[RTTTC]

One of my inverters goes to sleep at night, and I imagine - a lot of them do so.
Then I get bombarded with red alerts :)
It would be nice to have an option to not show these, or have alerts show up if inverter is unreachable for 24h or more..
Since PV meter is just for visual pleasure anyway :)
Any thoughts ?

[RTTTC]

Due to a large number of posts - ideas are quickly lost in the flood ..
@andig @naltatis @premultiply ?

[andig]

I think the problem is same as before. How does one know if an error is an expected error? Just swallow certain kinds of errors? Which ones (freely configurable)?

[RTTTC]

I think - yes - freely configurable.
Would it be possible to integrate some sort of 24h accumulation period/timer - and if the error lasts non-stop for more than 24h then show it ?
This would work nicely with pv meters (inverters) that sleep at night.

[VolkerK62]

what about time-based? No error between x and y o´clock

[mdkeil]

no error between sunset and sunrise.. with the need of set up latitude and longitude via ui/config (this infos you will need in the future for pv-forecast) .. evcc just can get this times for example from from https://sunrise-sunset.org/api.. or you use some of the existing go-implemenations.

[andig]

what about time-based? No error between x and y o´clock

Not under any condition. This is simply too fragile with little value add.

[naltatis]

I think this is mainly an issue in how we currently show errors to the user. It's not helpful for a user to see the same error message every update cycle or seeing an old notification of a network error that was fixed in the next update cycle. Such infos are interesting when debugging (we have log view for this now). But the raw info does not need to be pushed to the main UI.

The solution here is to introduce a per-entry/device (meter, charger, lp, tariff, ...) health state. If a device continuously produces errors or warnings we should notify about this in the main UI.

The overview of all device health would be shown in config UI. We already have tiles for each entry there.

Introducing such a health state is not an easy transition, but I'd suggest investing time into this concept rather than implementing custom muting rules.