In addition to the 65000 hard limit, there is an index space limit. Using long meta data (track title, artist name, etc.) may hit the space limit before 65000 tracks. The indexer continues until a limit is reached or all files are included. I use very short track file names in order to conserve index space. Track number is sufficient if each disc is stored in its own foulder.
You have virtually no control over processing order, but I notice that file shares are processed in order and a file share is fully processed before moving to the next. The indexer asks for a list of files to process. The presentation order can be chaotic as far as the human is concerned and tends to be associated with the physical order on the drive and this can depend when the file was added or updated on the drive. Displaying files by date may give you a hint about which files are at risk or which files are now beyond a limit.
Duplicate tracks can add to the confusion because only the first encounter will be included in the index and it’s not always obvious where the duplicate is stored. (A typo could store a duplicate in an unexpected folder.)
You may be able to somewhat manage things with separate shares, such as “dad”, “kid1”, “kid2”, “holiday1”, “holiday2”, etc. and only include currently needed shares.
