Maybe it’s unrelated but I get white screens when going into settings > addons, which are fixed with a refresh, so it “feels” like caching.
But why would it only be the app crashing when it goes to the background? I’ve got a device with the app always open that’s never had a problem, same with the browser.
Hopefully I’m not coming across as argumentative or anything, just trying to better understand how things work and troubleshoot the problem.
Lets just be clear on something here. The app is not crashing, just the HA frontend is unable to load. The error message you see is provided by the HA frontend, it is not anything app controlled. Its actually out of scope for the app at that point.
different devices will behave differently, are they both using the same Android System WebView version?
Only one of the other devices is an Android, but it’s a Kindle Fire and when I search the Play Store it indicates WebView isn’t even installed on it. I do know it’s the Play Store version of the HA app and I have everything set to auto-update, but I’m guessing it’s using some Kindle port of WebView?
The other always-on devices are just running the chromium browser in the Raspberry Pi OS (but on the original Surface tablet).
And then there’s my main Windows PC running it in Brave, but this isn’t always on (but I’ve had it open for extended sessions with no issues).
I’ll try opening it up on my phone (the one having issues with the app) in a browser to see if it gives me any issues. And in the time it took me to write all of this it’s giving me the same issue! It’s running 114.0.5735.60.
yea amazon does its own thing there and unfortunately updating webview is not an option there, you have to wait for amazon to push that update.
Does the WebView version I’m running yield any clues? As added context, it’s a fully up-to-date Pixel so it should both be as vanilla and as current as anything Android can be.
the only times ive seen issues with webview is when its outdated, I have not seen issues when its up to date. Personally I would start by disabling custom cards one by one to see if one is the culprit. You should also check open issues with each custom card as well in case someone already reported it.
In terms of investigating on your own you can try using Chrome Remote Debugging to see if it gives any clues
Same problem here
So I’ve found a few other tidbits.
1. I left one of my smallest dashboards with only 2 standard cards ‘open’ in the background. I got the same error but it recovered almost immediately.
2. I left one of my larger dashboards ‘open’ and got the same error, but purposefully let it sit for a while. Took maybe 30+ seconds but it also recovered eventually (guess I wasn’t ever patient enough before).
Edit: Seems I spoke too soon. Tried to recreate it all a few more times and everytime it acted a little different. Recovered a time or two, but not always and the size of the dashboard didn’t seem to have an impact.
@dshokouhi Do you think the potential problem custom card would need to be on the last viewed dashboard to cause the issue? That was the purpose of my initial test, as I’ve got quite a few custom cards installed and REALLY don’t want to have to try disabling them one-by-one…
custom cards actually get loaded in the browser as soon as you load the HA frontend IIRC they dont even have to be in use in any dashboard as long as they are added as a resource the code will load. In the past we had seen an issue where just loading card-mod would interfere with the apps more-info pop-up code.
alternatively you could always list your custom cards to see if others who have this issue may show the same custom cards. But I think if you just started the process of elimination since you can reproduce it quite easily it wont take as long, unless you had a hundred or some odd cards.
There’s a GitHub issue that’s been raised. Probably more likely to get attention from the developers than this forum, so I’d recommend posting there as well if you have the issue.
Thanks for the info/background.
Finally had some time to sit down and make a proper list of all my custom cards, etc. Going to follow your advice and remove them one at a time and wait for the error to pop up again (as a side-benefit, I think I have several installed that I don’t currently use, so I guess I can do some housekeeping while I’m at it too).
Plan is to remove each card one at a time and restart HA each time (but not the host machine). I assume this is the correct procedure to fully flush them from the system?
actually now that you found the frontend bug that matches your screenshot…check the comments to see if you see the same logs/experiences others do. Might be something else entirely different than my original assumption.
Unfortunately it’s the same general info as mentioned in this forum (i.e. nothing). Nothing to really go on; some folks seem to have something in their HA log and other don’t (such as myself).
Went through each of my custom cards and verified none of them are the cause. Hopefully something comes of the GitHub issue raised.
As I note in github, personally I have been able to significantly reduce the number of times I’m seeing the error on my mobile & tablet by reducing the number of events happening on my network to below 100 per minute. So until the root cause is fixed, for the short term I’m suggesting going into developer tools/events, subscribe to “*” and see how many events are happening in your network on average every minute. If there is something chattering away that’s not required, try stopping that and seeing if it helps at all. It might not help, but it’s worth a try.
Unfortunately this isn’t really doable for me. I have WELL over that many events under normal conditions, on the order of a couple thousand a minute.
Yowza. Ah well, fingers crossed something comes from the GitHub issue.
Having this also since the update to 2023.6 and only on Android, whenever i try to access on the iphone i see no issues (but i am not using the iphone on a regular basis).
Poco F1 - Android 13
My wife has the Poco x3 and doesn’t seem to have the issue, at least she didn’t tell me.
I found I had the issue on android, iphone and ipad - significantly reduced now I’ve been able to reduce the number of events being broadcast on the network but still not great. If you close the app after using it then you’ll never see the problem. If you leave it running in the background and/or your device goes to sleep then you will most likely see it when you try to go back to the app.