I'm running assets2, connecting to some S3 buckets. At the moment we're transitioning between one set of buckets and another, so I would expect our index to be twice as big as it will eventually be. But recent dumps of the database seem to be growing exponentially (it leapt from 10.4Mb to 40Mb in a week and there's an editorial team of one working on the build!) and I think it's the assets_index_data table - a snapshot this morning tells me that it has 3,045,452 rows.
I'd be surprised if we had 10,000 images in these collective buckets, so I can't figure out why that table is so gargantuan - surely it can't be right? It contains big groups of rows that all have the same session id - it's almost as if every time it indexes the buckets, it creates new rows rather than updating existing ones, leading to an enormously bloated table.
Of course I may be completely wrong and perhaps it needs to be this big, but it doesn't seem right and can't continue to grow this way!