Why 150 scroll contexts are created in elastic search?
I would have posted a comment instead, but can't due to low rep. Can this help ? https://www.elastic.co/guide/en/elasticsearch/reference/current/search-request-body.html#_clear_scroll_api
Also you can try and tweak that value in you cluster setting :search.max_open_scroll_context
I dont really get why you would have so many. But it had happend to me that clients were starting scrolls and not finalizing them, then this contexts would remain open until their expiration time of a default of 5 minutes. This would even kill the cluster if the scrolls had a high number of items matching.
In my case I had a scroll_current around 65 for a single scroll.
By testing on a test index with no tokenizers definition I had for each scroll exactly 1 scroll_current. Then by adding a definition of the index with some fields and an ngram tokenizer I've had back the high scroll_current.
My hypotesis is that the tokenizer has to consume more resources.