schema_version: 1
id: progressive-resource-growth-during-steady-workload
title: Progressive resource growth during steady workload
status: draft
owner: codex
source: jira
environment:
  profile: stable-dedicated
  cluster_id: '1379661944646413663'
  cluster_type: dedicated
  tidb_version: v8.1.1
time_window:
  incident_at: '2026-08-05T08:30:00Z'
  allowed:
  - id: reported_incident_window
    start: '2026-08-05T07:00:00Z'
    end: '2026-08-05T10:30:00Z'
prompt:
  text: 请分析集群 1379661944646413663 在 2026-08-05T07:00:00Z 到 2026-08-05T10:30:00Z 的资源异常。业务负载没有明显增长，但 TiDB 实例运行越久，内存和每条语句消耗的 CPU 越高，Go GC 开销也在上升；实例重启后指标会短暂恢复。请判断最可能的根因，给出证据链、影响范围和处理建议。
tags:
- resource-pressure
- cpu
- memory
case_tier: P0
cost_tier: medium
followup:
  scheduled_max_followups: 0
execution:
  acp_bash_timeout_s: 900