Aggregation performance

hello,

Can anyone please advice how to improve a query performance?

  • the bool filter takes 50ms.

  • filter dates: last 7 days

  • total search documents : 20075

  • total index documents : 79m

  • searched indices: 1

  • index primary shard size: 9.2gb

  • primary shards: 2

  • replica:1

  • storage tier: hot

thank you!

How much faster do you want it?

Usually I'm used to much better performance.
I'm trying to keep 2 seconds as target number

If your index is only 9.2GB in size, why have 2 primary shards? What is the hardware specification of the cluster? What is the full output of the cluster stats API?

I can decrease the shards via ilm, and do rollover..
attached below, thanks!

https://jsonblob.com/b6777ebd-cf36-11eb-a671-7b947ffb5d48

@Christian_Dahlqvist @warkolm any tips you can think of? Is 90 million too many documents per index?