知识卡片

普通读书笔记卡 · 1754.y.2

可观测性保存故障黑匣子

理解等级:中级

系统无法预见全部错误,应记录延迟、吞吐、队列、资源和业务指标,与预期基线比较;偏离时调查根因,不足以解释时补充观测。尤其要在失败当时保留上下文,因为罕见问题未必能再次重现。指标不是装饰面板,而是让下一轮设计从未知失败中学习的证据;报警应有解释和调查路径。

这张卡片看不懂?

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

    • 构建稳定系统 · 查看原书依据

      Most importantly, when a failure happens, we need to be able to capture as much information as possible to enable us to find the problem and fix it. Our system needs its own black box, like an airplane.

    参考来源

    - 位置:《Software Architecture and Decision-Making》第12章《Building Stable Systems》(来源项目:book-sources;源文件:extracted/Software Architecture and Decision-Making Leveraging Leadership, Technology, and Product Management to Build Great Products… (Srinath Perera)/_epub-src/ch12.xhtml) - 结论依据:摘录说明系统无法预见全部错误,应记录延迟、吞吐、队列、资源和业务指标,与预期基线比较;偏离时调查根因,不足以解释时补充观测。其余条件结合本章相邻例子与取舍说明,数字示例不作当前性能承诺。 - 原始内容:Most importantly, when a failure happens, we need to be able to capture as much information as possible to enable us to find the problem and fix it. Our system needs its own black box, like an airplane.