Lovelace ui add card panel freeze / unresponsive for long time

Hey everyone, i’ve encountered a problem i haven’t been able to diagnose and solve on my own.

i recently moved from yaml editing to the lovelace ui editor.
everything works fine, two weeks ago i’ve noticed something strage.
when ever opening the ui editor to add new card (on what ever tab i choose, clicking on the plus button to add new card)
the home assistant is stuck and cant be moved, it looks like some component making it unresponsive and after i wait a minute or two it come back to life, or i close the tab and open hass again.

in different instance of ha i have for my family this issue doesn’t reaccure. i’ve tested it with few different browsers and pcs. so it’s defenitly something with maybe a custom component i’ve added but i can’t figure out which.

did anyone encountered an issue similiar to me? search results came up with nothing i could find.

my configuration: running latests HA docker container (being updated automaticly using watchtower) so it’s not related to specific version of HA.
running on xeon cpu with enough power so can’t connect it to power related.

thanks, have a great day.

no one encountered this problem?

or has any clue how to attack it to find the cause?

I am brand new to HA, and just successfully added my first sensors and cards to my Overview screen.

Now, just this morning, I am encountering the exact behavior you describe. I end up closing the browser tab and re-starting.

I will now hope to begin yaml UI editing until some solution is uncovered.

Running 0.109 on Vbox Ubuntu machine on a Win10 NUC. I did click the update to 0.110 button earlier today but it does not seem to have taken, even with restarting.

FYI . . .

I’ve been using HA for more then a year now.
so i’m not sure what’s the connection between both cases.
interesting how happend to both of us. couldn’t find referecne so far for more people having that issue.

btw didn’t solve it up untill now, i’m replicating existing cards and then edit them or more to the tab i want…
i dont the lovelace anyways because i’m trying to automate most of my cases. but monitoring pages are always a work in progress so i want the ui editing to work as expected.

I have the exact same problem.
It has been like this for some weeks I think.
Browser freezes and I have to closed the tab. It doesn’t help to wait.

I have tried on different browsers and computers.

Running HA on Ubuntu server with docker.

Maybe not the problem you’re facing, but I have found that this can be extremely slow as it loads the card preview with real-life entities, so you can end up with a case where it shows an example of a history graph loaded with data from sensor.time or something equally (or more) data-intensive.
Multiply that with all the cards it wants to show you and you end up with a sloooow loading page.
I really wish it could show dummy data instead.

I had the same problem. As Skeletorjus writes, it takes a live sensor. I added a snmp sensor for network traffic with that sensor enabled in config lovelace add card freeze.
When I removed the sensor, the UI worked again.

@Skeletorjus @mats

that’s what i think is happening, if i wait enought time (even when chrome said the page is stuck, ask me to wait or kill, i tell him to wait, eventually 2-5 min, it will load and will be responsive again)

the question is how do i figure out which sensor/entity make this happen, and what can i do if i dont want to delete this sensor.

like @Skeletorjus says it would be helpfull to be able to put dummy entites there

@mats

that was the answer that pointed me exactly to the right direction
i had only one snmp sensor, the moment i removed it, the add new card panel return to work again!

it’s actually make sense, this sensor was changing every second and has different value, to display it in history graph might cause it to load for a long time.

thanks!
hope it will help others

1 Like

A workaround is to simply exclude the offending entities from the recorder.

1 Like

I must be a victim of the frequently updating serial sensor I recently added.

How do I remove this from the recorder?

Thanks.

Never mind, I think I found it.

Recorder: etc.

Same issue, how do I got about trying to track down which sensor is causing the issue? I don’t have any new sensors yet mine just started doing it as well.

edit: Mine was because I changed IP’s and Brave-Browser had re-enabled ‘Shields-Up’ mode.

I am having this same issue except I do not have any live sensors that I know of. Can anyone point me in the right direction on how to find out what is causing this?

As a follow up in case anyone else runs into this issue mine seems to be resolved for the time being. I restarted the HA host (I am running in a VM) and had an error stating my default config was invalid and history or logbook could not start. The fix for that was to delete the db file. Here is the post that explains that: https://github.com/home-assistant/core/issues/7012
After doing this and rebooting the host again, the card picker loads fine.

My guess is that the db was corrupt and causing issues with the card picker and other things. I am going to keep monitoring to see if it happens again, but for now it has been running a couple of hours with no issues.

I had this problem as well, and was able to solve it after reading this thread.

To find the issue I was looking at all the example cards when adding a new Lovelace card, and looked for anyone with complex data.

For me it was the “sensor.time” sensor which was populating the History Graph example like this:
history-graph

So I removed it from the recorder like this:
(I removed some other sensors as well, as neither of them makes sense to record).

recorder:
  exclude:
    entities:
      - sensor.time
      - sensor.time_date
      - sensor.date

After waiting 24 hours for the sensor to clear up from the previous data, the issue was solved. :slight_smile:

Thank you all for the tips and solutions in this thread.

1 Like

This is the top Google response for “lovelace unresponsive home assistant” so I’m just posting my experience in here.

I had a perfectly working Lovelace dashboard for ~6 months, and then it just came unresponsive. No changes to sensors/etc, all stayed working/connected.

I would login, and it would work for 2-3 seconds, then it would hang for several minutes and eventually work. I tried updating to latest versions, nothing worked.

I ended up deleting all of my Lovelace dashboards, and recreating them. This resolved my issue. Not ideal, but it got me working again.

you can see the history graph loading state data... when you bring up the add card modal.

I had added a bunch of date_time sensors so added a few more to exclude list.

This issue seems BAD BAD, a serious js async/crud mistake. Never hang the entire browser waiting for an async operation (like a crud operation) and why preload ALL the data for all entities when the user has not even chosen the history card let alone chosen the entity. On top of that the backend should at least warn when a sensor is recording a huge amount of data

Seriously I would have never allowed a frontend with this issue to go into release.


sensor:
  - platform: time_date
    display_options:
      - 'time'
      - 'date'
      - 'date_time'
      - 'date_time_iso'
      - 'time_date'
      - 'time_utc'
      - 'beat'
recorder:
  exclude:
    entities:
      - sensor.time
      - sensor.time_date
      - sensor.date   
      - sensor.date_time_iso       
      - sensor.date_time_iso   
      - sensor.date_time_iso  

now I just have to wait for db to clear :(.

Nah, it’s really not as bad as you describe here.
I do agree it’s an annoyance, and especially before I found the solution.

But it does not pre-load all the data for all entities; It just picks one to use live data for the preview. And if that sensor does contain a lot of data this issue will happen.
That is also why this issue happens even if you have not chosen the history card.

The beauty of Home Assistant is that it’s crazy powerful, with a lot of possibilities to choose from.
So when you add sensors you also have to decide if you want to record them or not.

Sure there are room for improvement, you could probably check GitHub and open an issue if it’s not there already.