I really appreciate the default auto-generating “Overview” lovelace UI, as when devices and entities are added I can be confident that this dash will be updated at least for an initial play. In fact I’ve been using this as my only dashboard since I started.
But as the number of devices on my HA instance grows, it’s becoming a little cumbersome, especially via the mobile app. I’m assuming this means that I’ve possibly I’ve outgrown it, which is fair enough, but does that imply automatic generation always has a shelf life? Is there a way to parameterise it?
What I would love is to have separate, individually controlled, dashboards created per room as an opinionated, yet logical, way to partition devices. As I add devices, or move them around physically, they would also jump to the relevant dashboard. If rooms are too opinionated then perhaps a tag or other group.
Failing that, and if I have to ultimately take control of individual dashboards, I’d want a way to transfer cards to those, while maintaining a list of yet to be assigned automatically generated cards. That way I can have a view of new but uncatalogued devices to add to my real dashboards.
I suppose this question is actually asking about workflow. For those who have outgrown a monolithic auto dashboard, how do you update the relevant custom dashes with the same cards you would have gotten automatically?
I’m pretty new to HA but I’m having the same feelings round my ever growing lovelace dashboard. When talking to my HA mentor, he suggests that once I’m comfortable with HASSIO, integrating systems and the like, that I should make the switch over to YAML and I think you might be there as well.
For me, I don’t even WANT a HA dashboard. I just want to bring everything INTO HA and then I’m using Node RED to set up true ‘automations’ that just ‘do what they are supposed to’. Example: I used to think that I wanted a page for my wife to go to prior to travelling to our 2nd home where she could log in, press a button saying “we’re coming” and then the house would be ‘ready’ when we arrived.
Now I realize that she isn’t gonna do that (nor should she have to) so I automated the whole thing to check my google calendar and if it sees an entry with a specific name it just adjusts all the home settings for our arrival without ANY interaction from me or my wife. There’s no buttons to press, no apps to download, nothing. All she or I have to do is put on our schedule that we’re gonna be there on a certain day and HA + Node RED handles it all for us.
If you’re still set on having a dashboard, check into using YAML for your instance instead of lovelace. Its the next logical step for you.
It’s not that I’m resistant to YAML it up (although I think the UI editor is plenty) - my previous HA instance only had custom cards. It’s more about completion and discovery.
A single new integration can introduce multiple devices, each with multiple entities. Without A-Bing the device/entity registries it’s sometimes difficult to track what you’ve consumed and what you haven’t.
But you make a fair point about the UI in general - maybe I don’t want what I think I want, and should keep the default UI as the exhaustive palette I build my (ultimately more sparse) manual UIs from.
I too am struggling with device, entity, component, group, area UI overload…
I have started to use the HA command line tool ‘hass-cli’ to query Home Assistant. Trying to find the optimal use of groups, areas and other attributes to organize stuff so I can logically use them in UI. Have a look at this tool:
For example, a command line sequence on MacOS to give me a spreadsheet of temperature sensors. You can use the output of hass-cli as a starting place for organizing your UI items. I am from the ‘old’ days before Lovelace, so I from there I moved to Lovelace YAML config of everything. This too I am now wondering if I need to scrap and go with the UI based starting point to configure Lovelace.
hass-cli --table-format=tsv --columns="entity_id, state, UNITS=attributes.unit_of_measurement, attributes.friendly_name" state list '^sensor.*temp.*' > /tmp/tmp.csv | open -a /Applications/Numbers.app '/tmp/tmp.csv'
This is what I’m doing. I let EVERYTHING hit the lovelace default. This provides lots of insights as you’ll have sensors you didn’t necessarily ‘think’ you’d have off of devices you didn’t think would have them! It just helps me to ‘visualize’ what things I’ve got in my head when I can see the devices.
Let me know when you wake up in a cold sweat and start sketching Node RED flows on a notepad next to your bed…
