Hello @Kishore Nallan! My Scenario: Hardware: 3 Nodes - 32gb - 8 CPU cores Previously I had all data in 1 collection (12 million skus+prices). Each sku has many prices/sellers. This was a "fat" collection and it was taking about 6-7 seconds for similar query. Then I changed the data modeling to split skus and prices in 2 different collections. Using "joins" the timing is now around 500-700ms. 2 collections: skus (about 120k skus) and prices (about 12 million prices) This is a very simplified version of one of the most common query I have here:

{ "searches": [{
"collection": "skus",
"query_by": "sku_name,product_name",
"q": "*",
"per_page": 10,
"include_fields": "product_id,sku_id,seller_id,sku_name,product_name",
"filter_by": "$offers(offer_active:=true && clusterId:=['null','abd-1234'] && paymentFormId:!='null') && seller_id:=['670972463814f3334d650c74','66a9981f9e6d8686ba64a3f3']"
}]}
"search_time_ms": 417

And I have this scenario which is very weird. This query should be very fast right? it's very simple and I just want to get 10 active offers.

{ "searches": [{ "collection": "offers", "q": "*", "per_page": 10, "filter_by": "offer_active:true" }]}
"search_time_ms": 1974