← 返回刚才阅读的位置

知识卡片

问题卡 · 4545.e.3.c

卸载与降级依据不同信号

理解等级:中级

配额限流按预设速率约束调用者,负载卸载根据CPU、内存或队列等饱和信号主动丢弃部分请求。优雅降级则减少每个请求的工作,例如返回可接受的缓存或更便宜算法。三者可配合,但必须事先明确哪些结果允许缩水,不能把降低正确性当作普适的降级。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • Resilience · 查看原书依据

      Load shedding is a technique used to predict when a server is approaching that saturation point and mitigate it by dropping some proportion of traffic in a controlled fashion. Ideally, this will prevent the server from overloading and failing health checks, serving with high latency, or just collapsing in a graceless, uncontrolled failure. Unlike quota-based

    参考来源

    - 位置:《Cloud Native Go》第09章《Resilience》(来源项目book-sources;源文件:extracted/Cloud Native Go (Matthew A. Titmus)/_epub-src/OEBPS/ch09.xhtml) - 结论依据:配额限流按预设速率约束调用者,负载卸载根据CPU、内存或队列等饱和信号主动丢弃部分请求。原文相应定义、示例或限制为此结论提供依据;补充纠错与版本边界在正文就近标明。 - 原始内容:Load shedding is a technique used to predict when a server is approaching that saturation point and mitigate it by dropping some proportion of traffic in a controlled fashion. Ideally, this will prevent the server from overloading and failing health checks, serving with high latency, or just collapsing in a graceless, uncontrolled failure. Unlike quota-based