Continus writting will cause segment merge, therefore cause filter cache invalid, is it way to monitor this?

[makeyang]

my cluster cover a search request with a lot of new data indexed, then the search performce will drop greatly time by time. is there a way to monitor the invalid of the filter cache caused by segment merge?

[warkolm (Mark Walkom)]

Any segment merges invalidates filter caches, as they are on a segment level.