2023.2: How can I Assist?

Cloud native meant as an approach to doing software development, not as ‘give-all-your-data-to-the-cloud-overlords’. Like in building a factory for software development by automating as much as possible along the pipeline. Making sure you go through the necessary tests at every step. But yeah, totally agree, keep your data as private as possible.

It reminds me of the time when HA decided to rename the old Hassio install method to just plain “Home Assistant” even tho there were three other ways to install it.

“What version of Home Assistant do you run?”

“Home Assistant”

“yes, but which one?”

“Home Assistant”

etc. etc

:grinning_face_with_smiling_eyes:

It is not a workaround. A workaround is a (usually undocumented) way to get around a software fault. Disabling updates is the way it is supposed to work, and does work.

1 Like

Don’t get the point, what’s the issue with searching or talking about „Home Assistant OS“ or „Home Assistant Assist“?

Never had problems googling those terms when using quotes around „Home Assistant“

Hi team, sorry for my double post But I think this is also the right place to ask here.

I may not swear. But F#, 5th time in a week all my devices where “out-of-date”. 42 in total. Before I asked (in a WTH) if I could opt-out devices in the esphome interface not to update.

But now updates are even more prominent in my home assistant interface.

42 updates available.

And I feel there’s no good or bad here. I believe being up to data is always better than lacking behind, especially in the IoT world. But 5x42 updates in 6 days? Skipping updates or disabling notifications is not a valid way. Updating all 42 with a 1-2 day interval evenly not.

What is the proper way to manage this HomeAsstant (and ESPhome) team?

I was not getting any update entities from ESPHOME and found a setting in the ESPHOME add-on I had switched off:

Home Assistant Dashboard Integration

Once I switched that on, I got lots of update entities as mentioned here.

What else do I need that enabled for apart from updates or can I leave it switched off ?

The best way I’ve seen is in post 419, above.

If I understand correctly, the latest updates of HA and/or ESPHome integrated the two more tightly. Part of that integration was adding a bunch of entities to HA. Some of these are update entities. You can see them in your entities list in Settings / Devices and Services / Entities. I had not noticed these before, but apparently this is how HA knows a version of something has changed, which triggers the notification “bubble” in the UI.

Maybe adding an update entity for every ESP device was helpful for someone. I have yet to see that explained. But at any rate, most active ESP users will probably find it problematic, for a number of reasons which have been well aired on this forum. Fortunately, there is a way to avoid (i.e.; “work around”) those problems. Just disable those update entities, right from that same Entities page.

1 Like

Just look up a few posts. It’s very easy to turn off update entities.

It’s firmware tracking. It’s built the same way as all devices that have firmware updates. E.g. zwave, zigbee, UniFi, etc. It was an update to the esphome integration.

FYI: update entities were added in 2022.4.

Hi, thanks and yes seen that. But disabling updates entities is not managing, is it? it’s just ignoring updates. Is there not any options to “forget” this update and wait for the next or other?

You can write an automation to skip specific updates with the skip update service. Works like any other service, just supply a list of update entities that you want to skip to it.

One gets more n more curious, what’s this " update annoyance " is all about , is their any specific reason that ESP-devices needs/is/gets so frequent updates ? , is it bug-fixes, feature updates or what/ why does it sounds like ESP-s has so frequent updates ?

I don’t have ESP-devices(what i know of), thou bunch of different brands in WIFI and Zigbee devices, but the “update” frequency , is very far from what current “Buzz” about ESP indicates

Esphome just updates the software frequently as they have a full time developer.

Ahh , so it’s not the various devices “firmware” in specific, but the platforms ?,

Esphome is the firmware.

@Guillaume_Corgnet Did you figure this out? I have the same issue.

Updated to 2023.2.5 from 2023.2.0 Somehow all my Google Nest cast dashboards are “damaged”. Half of the entities are missing (consisting out of a lot of mushroomcards in a panel setup) I’ve restored a backup of last week when there was no trouble but that did not help. Few others see the same behaviour (thread). It looks like other dashboards that are not “hand build” do work. Anybody else experiencing this behaviour or maybe even have a solution. Needless to say that the family is not amused…

Before


After

There have been some minor adjustments to the dash but you’ll get the point…

3 Likes

Hi, I’m have the same issue. Cards show fine within HA dashboards, no console errors - but when casting to Nest Hub, or Android TV - the same cards are missing. Up until reading @DBM’s post I thought I must have messed something up - but as I’m only just in the process of setting up a Cast dashboard - I didn’t have a ‘before’ to compare it to.

I’m seeing some Mushroom chips missing, but not all / Sensor card / Mushroom cards / Mushroom lights - so my initial thought was that it was custom cards, but the Mini graph cards work fine?!

If it helps, I tried the CATT method, which shows all cards, but I’m trying to stick to the inbuilt HA cast method. I’m happy to try anything, if anyone has any suggestions.

Combined image, as it’s my first post so still restricted:

I use a sensor group to sum up my energy meters for a total energy consumption.
Unfortunately, when setting up the sensor via the new Group feature in the UI, there is no state_class provided to the sensor and thus, it cannot be assigned for measuring the consumption in the energy dashboard. Brings me back to configuring the sensor via configuration.yaml.

Am I missing out something here or is the state_class missing from the UI configuration on purpose?

There is a ticket open on the problem when casting: