Some observations on the schema update, just as a data point.
Before update I modified the LXC:
Hello,
Is it possible to have 3 centered cards (1/3; 1/3; 1/3)?
layout_options:
grid_columns: 1 (or 2,3,4)
2 is too large and 1 too small⌠
So having issue with Rachio Integration. I can control the switches, but I get no state updates. Unfortunately, my certificate expired at the same time I upgraded so not sure if what I am seeing is related to the update to 2024.8 or the certificate change. I am going to see if there is a Rachio integration message board but was wondering if this was seen by anyone else. To be clear the integration appears to work it is just not updating the any of the states in HA.
The only changes made to that integration were related to the new calendar entity. It doesnât appear to touch any other aspect of that integration.
Yeah, that is what I saw as well. I can see the packets to and from HA to the api.rach.io address when I perform action in HA however, I do not see any packets when I make a change in the app. It will turn on the switches but then they go back to their previous state and do not reflect the current state and if I change in the rachio app there is no update. I am really wishing I did not do both the certificate and update on the same day. Thanks!
Are you sure this isnât a caching issue? Clear your cache and refresh the page.
iâm here to complain about this. as I edit my automations and scripts after this release, the service element has been replaced with action as expected⌠but it moves the action element to the bottom of the json string, which is very annoying.
i manually had to edit my automations.yaml and scripts.yaml file with a search/replace command to avoid this, which is what I had expected this update to be doing in the first place. as someone who uses the interface yaml editor, the only thing i want to come before action is alias (if present) for readability.
I am in the same situation. Running MariaDB on Synology NAS 720+, it took about 17 hours with database over 20GB and failed three times in a row. I do no have stronger server, only desktop computer. This is really a failure of the update if it is only dependent on the computational power of the server.
No it is a failure to heed the often given advice to not store excessive amounts of data in the default database. If you want to store state data for the long term use a time series database like InfluxDB.
I have influxdb running as well and thatâs taking still the data but for long term statistics, I need this one. So what is your recommendation, keep only 30 days instead of one year? Long term statistics are kept anyway?
Yes 30 days at the most. LTS are kept forever no matter what purge days setting you have.
Where can I find that advise, and how to optimize things on my HA installation?
Pretty interestingâŚ
Can I drop just a part of the database and keep only the LTS? Is it the row âstatisticsâ for LTS?
Maybe this information should be on the main webpage then and not scatter throughout the forum, which is what I guess youâre referring to. The only thing mentioned in the docs is that if you want to have more real data then you should have more disk space. And thatâs it.
I donât mind having 15GB+ DB. HA gives this option so itâs not a failure, on either side actually. Especially since there was no so advanced recorder/history not so long ago and like I said - there is no obvious suggestion to make it like youâre suggestingâŚ
You can assume that data is sacred here 
and everyone is afraid of loosing it. Thatâs why itâs so important to have a proper guide to explain everything (if there is one - I couldnât find it), including fact that data hoarders would rather choose external machine, with probably something different than SQLite.
I tried once and Frenck killed it with absurd defensive arguments that my proposal is beyond the need of the documentation. There were dozens of people discussing it on GutHub and it was a problem to add FIVE words to ensure nobody spends days on deducing something that is self-evident to the people working on HA on a daily basis but is missed in the step-by-step instructions. Sorry, this should change.
Itâs awesome that you can feedback something not knowing that there is something to feedback about, right? 
Actually, your suggestion about failing to realize this hidden knowledge could be summarize the same way:
We should really not assume that everyone sits on the forum getting knowledge from each post that happens here. I for one, had no idea that there is some difference in the recorder. And thatâs because when I first configure it, there was no options like that and description was: you loose data when no hoarding it. Simple as that.
exactly, we could be given a recommendation that database update is going to happen and that we should consider lowering days for purging the database. LTS would remain and everybody would be happy.
Thatâs fine, I have no idea what you are on about so 
Cache was cleared, also see no change in state as reported by Developer Tools/States and no historical changes after update. I do see all the previous state changes though. I have now found that I have at least 1 automation not properly working and the widgets on my phone non longer update unless I have the app open. They show stale and normally when they are in this state, I click on them they update. Now they will not update until I open the app and then then when I click, they update right away. The automation that is failing is also tied to the phone. I am going to see if the automation works if I have the app open. I also had some issues using some esp32 smart speakers where one turned on an incorrect light and also has returned a device state was on when I was in the room, and it was off. I rebooted both of those devices last night and they worked after that so hopefully this glitch in the matrix will be resolved. I already rebooted the phone and the Rachio with no changes. So just found the assist on the phone does not work unless the app is open as well.
