Terms Aggregation Partitioning + filter buckets

[Nam_Nguyen]

Hey ,

Context

I am currently struggling to create a proper query for the following use case.
Let's say I wanted to create a group-by aggregation for e.g. term app:.*, I would do the following:

          "terms": {
            "field": "tags",
            "include": "app:.*",
            "size": 300000
          }

This would give me all buckets which matches the regex.

Why

I have some high cardinality fields + nested terms aggregations which can trip the circuit breaker

Questions

I have the following questions:

How can I combine above with partitioning? AFAICT it is not possible to filter what kind of buckets we will create and therefore for every tag a bucket is created. Is there a way to filter it like above?
Something like this?

          "terms": {
            "field": "tags",
            "include_regexp": "app:.*",
"include":{"partition":0, "num_partitions": 100}
            "size": 300000
          } 

Does ES recomputate the partitions each time we query them e.g. partition:1 and the next query partition:3 or does it somehow cache the values?

Thanks