yugabyte/yugabyte-db

[DocDB] Metrics for vector compaction and flush

Open

#28,578 创建于 2025年9月15日

在 GitHub 查看
 (1 评论) (0 反应) (1 负责人)C (8,229 star) (1,003 fork)batch import
area/docdbgood first issuekind/enhancementpriority/medium

描述

Jira Link: DB-18272

Description

Add metrics for vector compaction and flush related activities

Regular tables update the following compaction metrics:

  • rocksdb_compact_read_bytes
  • rocksdb_compact_write_bytes
  • compaction_tasks
  • rocksdb_compaction_times_micros_sum
  • rocksdb_compaction_times_micros_count

Need to think through the below counts/ Introduce specific counters for vector chunks

  • rocksdb_numfiles_in_singlecompaction_count
  • rocksdb_numfiles_in_singlecompaction_sum

Some of them are relevant to vector indexes. It would be good to update them for vector compactions as well. At the same time, we should add new metrics that describe vector compactions better.

Issue Type

kind/enhancement

Warning: Please confirm that this issue does not contain any sensitive information

  • I confirm this issue does not contain any sensitive information.

贡献者指南

技术栈
ccpp
领域
databaseperformanceobservability
议题类型
feature
难度面向新贡献者的预计实现难度，1 表示很小改动，5 表示专家级工作。
4
预计时间有经验贡献者完成调查、实现、测试并准备 pull request 的粗略时间范围。
3-5 days
活动状态议题当前的可参与程度：新鲜、活跃、陈旧、阻塞或等待维护者输入。
blocked
清晰度议题是否清楚说明期望改动、验收标准和下一步。
mostly clear
前置要求
Understanding of compaction in distributed DBFamiliarity with RocksDB metrics
新手友好度1-100 的估计分数，表示该议题对首次贡献者的友好程度。
20
研究方向
Start by examining existing compaction metrics in `rocksdb compact *` files (likely in `src/yb/rocksdb/`). Look at how vector indexes are implemented in `src/yb/docdb/vector index/`. The issue requests adding metrics similar to regular tables but for vector compaction. Consider both updating existing counters where applicable and introducing new counters specific to vector chunks. Review comments on the Jira link for additional context.