[Star Tree] Evaluate the need of equivalent deferrable bucket collector

[sandeshkr419]

Is your feature request related to a problem? Please describe

Follow-up from a bug fix: #18354 (review)

Presently, whenever a search query is resolvable via star-tree, we don't wrap the collector in a DeferringBucketCollector and pre-compute the results in the same manner irrespective of BFS/DFS sub-aggregation mode.

This issue is to track use-cases with high cardinality buckets to see if there is an implicit need (running out of memory maybe?) to have an analogous wrapper for StarTreeBucketCollector similar to DeferringBucketCollector (for non-star-tree cases).

We have a decent experimentation with nested aggregations already with benchmarking results from #18048 where we did not run into any such issues. This issue will aim to gather more data from dedicated benchmarks to test this out and carve out any action items that may come next.

Describe the solution you'd like

Evaluate need for a deferrable bucket collector in star-tree

Related component

No response

Describe alternatives you've considered

No response

Additional context

No response