opensearch-project/OpenSearch

[Enhancement] Histogram aggregation should initialize Ranges lazily

Open

#18,241 建立於 2025年5月8日

在 GitHub 查看
 (13 留言) (0 反應) (1 負責人)Java (8,123 star) (1,505 fork)batch import
Search:Aggregationsenhancementgood first issue

描述

Is your feature request related to a problem? Please describe

Currently, optimized histogram aggregation using multi range traversal initialized Ranges with all buckets eagerly. While for individual request, memory implications are small, it can starting adding up across all requests on a node.

Describe the solution you'd like

Given we process buckets one at a time, they can be initialized lazily avoiding unnecessary memory allocations. This should also allow to raise the number of buckets limit from 3k to 10k or more

Related component

No response

Describe alternatives you've considered

No response

Additional context

No response

貢獻者指南

技術棧
java
領域
backendperformancesearch
議題類型
performance
難度面向新貢獻者的預計實作難度，1 表示很小改動，5 表示專家級工作。
3
預計時間有經驗貢獻者完成調查、實作、測試並準備 pull request 的粗略時間範圍。
1-2 days
活動狀態議題目前的可參與程度：新鮮、活躍、陳舊、阻塞或等待維護者輸入。
blocked
清晰度議題是否清楚說明預期改動、驗收標準和下一步。
clear
前置要求
JavaOpenSearch aggregation internalslazy initializationmemory management
新手友善度1-100 的估計分數，表示該議題對首次貢獻者的友善程度。
40
研究方向
The issue proposes lazy initialization of the Ranges object in histogram aggregation to reduce memory allocations. The implementation is likely in the aggregation module of OpenSearch, specifically in classes related to histogram aggregation. Review the current code where Ranges are eagerly created for all buckets and modify it to create Ranges only when needed, following the multi range traversal pattern described. Check issue comments for any specific file references or maintainer guidance on the preferred approach.