i'm working with a massive data import right now

it makes me think that keeping the raw events separate from the filter search index would be a massive performance boost and doubly so for bulk import, but generally for heavy write load

a blossom store could be a perfect place to put events