知识卡片
告警必须兼顾切片与可行动
阈值应从正常用户体验和服务目标出发,并持续依据误报、漏报与值班负担调整。全局1%错误率可能掩盖某个1%用户群体完全不可用,需对关键群体或异常变化另作检测。不是每个指标都要叫醒人,规则应指向可处理问题,异常检测也不是万能替代品。
这张卡片看不懂?
这张卡片的反馈已满 10 条,暂时无法继续提交。
已有反馈 · 0 / 10 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- Chapter 14. Monitoring and Observability Patterns · 查看原书依据
an error that causes a 100% outage for a small (1%) subset of your users. From a static alerting perspective, your system is functioning within bounds, but for those poor users in that subset your service is completely down. Static alerting assumes that all metrics are uniform
参考来源
- 位置:《Designing Distributed Systems, 2nd Edition》第14章《Chapter 14. Monitoring and Observability Patterns》(来源项目 book-sources;源文件:extracted/Designing Distributed Systems, 2nd Edition/_epub-src/OEBPS/ch14.html)
- 结论依据:原文小群体全故障例子与值班告警疲劳说明聚合局限和调整必要性。
- 原始内容:an error that causes a 100% outage for a small (1%) subset of your users. From a static alerting perspective, your system is functioning within bounds, but for those poor users in that subset your service is completely down. Static alerting assumes that all metrics are uniform