Describe the bug

Problem
The reduce phase on coordinator nodes was a bottleneck, causing high CPU usage and latency, especially with dedicated coordinator nodes required by customer policy.

*Solution
Optimized the reduce logic by:
* Using a merge-sort approach for key-ordered aggregations to efficiently combine buckets from multiple shards. This is useful for high-cardinality cases with keys having 2 or more duplicates ultimately resulting in less number of comparisons to perform reduce merge and fetching topN.
* Applying quickselect instead of PriorityQueue when the final size was smaller than the bucket count, avoiding heap overhead.

Changes:
rishabhmaurya@3c25140

Related component

Search:Performance

Additional Details

Plugins
NA

Screenshots
NA

Host/Environment (please complete the following information):

• OS: [e.g. iOS]
• Version [e.g. 22]

Additional context
Add any other context about the problem here.