How to index/search data in very-large subset of documents?

Elastic Stack Elasticsearch
1

Hi - this might be a newbie question but I could not find a related answer in my search.

The use-case contains a lot of documents (100 million+) each tagged with a unique case number. The number of case numbers are currently 100 thousand or so. Thus, each case number can have 1000+ documents easily.

We want to do a search within a group of case numbers for a given agent - say at the max say 2000 case numbers.

How do we model this:

  1. Should we fire an IN query in case numbers during Search.

  2. Or, should we have one document for a case number - and update the same search-document when a new user-document is added against that case number.

Thanks in advance.