I’m backing up 8 TB of disk images in a single backup set. Works fine, 64 mebibyte deduplication blocks and 4 gibibyte volume files.
Probably smaller dedupe blocks would increase effectiveness of deduplication, but in this particular case it doesn’t matter so much. Most important thing is to have a copy of the data with version history.
With older versions this set used to corrupt often, but it has been working fine for at least half an year so far. (Backup versions are being created daily).
Another large backup set is the backup of all of the normal duplicati backups. Yet these massive jobs are excluded from that one.