The Universal Scraper for Music Albums (the subject of this thread) is based upon identifying the album using Musicbrainz, and so access to their servers is essential.
To outline how it works: If music files are tagged with Musicbrainz Album Id (
release id), then idenitifaction is obvious, otherwise it does a lookup on album and artist. If a match is found (and 1st release will be taken if there are many) then it fetches artist credits and track data, which if you have "Prefer online info" enabled could be used to sync your lib with MB. Then it uses the
release group id (something we don't currently fetch from tag but could) returned by MB to access TADB, and fanart.tv for other info and art. If present it uses any allmusic url returned from MB (they have links) to access data from allmusic.com
However could there be other music scrapers, that do not use MB, just up to someone to write them (in Python now, the scraper API has been extened to allow that). I think there is a TADB sepcific one, but I don't know if that uses just album title and artist, or has a quiet MB lookup tucked inside. But TADB also runs out of server resources too.
I think we could improve things by reducing the requests Kodi makes to MB (or any other server). For example if it had been easy (i.e. automatically happens) to store links to those missing thumbnails of yours, Kodi would not be trying to find them all again. Likewise we could store the release and release-group IDs. I am working on both, slowly because the scraper "regexp parsing of xml" side of things is a mystery (@
olympia can you help?)
Otherwise a donation of more server power to the global comminity would help I guess - throw more hardware at it - but demand can be a bottomless pit.
EDIT: Just to add that IMO using MB to identify albums is a good idea, album title and artist is often not enough but if that is all we have then MB is the place to find it accurately