You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The library items are retrieved by the frontend when you open the panel all-at-once.
So all Artists are loaded in one batch, All Albums etc.
For people with larger collections this comes with a performance penalty:
There's a real large chunk of data being moved from db --> MA --> HA --> websocket --> frontend
It takes a while before the frontend is useable, you have to wait until the data is precached into the "store" of the frontend.
The HA executor is busy creating large chunks of data, I believe @bdraco has recently seen a profiler report showing this.
Solution here would be to move the item listings around as smaller chunks.
Like prefetch the first 250 items and request more if you scroll around the listings (or slowly prefetch them if you are navigating)
Not a very large job to adjust but needs to be done properly. Too much to make it into the 2022.6 release tomorrow.
This task is a reminder.
There are several issues reported directly or more indirect belonging to this issue.
@bdraco thanks, very helpful! Just as I suspected the load is in getting the library items all at once (e.g. tracks, albums etc).
this is also going to have a huge performance benefit of your orjson project ;-)
This is now implemented. Only small chunks at once are grabbed. This will save a huge performance hit and it will be even better once the orjson support hits HA core
The library items are retrieved by the frontend when you open the panel all-at-once.
So all Artists are loaded in one batch, All Albums etc.
For people with larger collections this comes with a performance penalty:
Solution here would be to move the item listings around as smaller chunks.
Like prefetch the first 250 items and request more if you scroll around the listings (or slowly prefetch them if you are navigating)
Not a very large job to adjust but needs to be done properly. Too much to make it into the 2022.6 release tomorrow.
This task is a reminder.
There are several issues reported directly or more indirect belonging to this issue.
The text was updated successfully, but these errors were encountered: