Cross posting from stackoverflow
I have gone through this link
It is not clear to me what does indices.store.throttle.max_bytes_per_sec actually do? Is it like in one second if segments that are getting merged are having size more than 20 MB then it will throttle the indexing request?
Can someone also explain me how does increasing the value of max_bytes_per_sec effects disk i/o? It appears to me increasing the value will lead to more disk I/O as now more segments can be merged ? Is it correct?
Can someone let me know when does merge process kicks in? Is it like when the number of segments greater than x then merge process starts?