<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>成本优化 - 标签 - Shengxu · 云架构 &amp; DevOps</title><link>https://shengxu.pages.dev/tags/%E6%88%90%E6%9C%AC%E4%BC%98%E5%8C%96/</link><description>Shengxu 的云架构 &amp; DevOps 技术博客：Kubernetes、Cilium、可观测性、LLM 基础设施、Agent 工程化等。</description><generator>Hugo 0.153.2 &amp; FixIt v0.4.0-alpha.3-20251225101113-8ffb9a95</generator><language>zh-CN</language><lastBuildDate>Sun, 04 Jan 2026 17:00:00 +0800</lastBuildDate><atom:link href="https://shengxu.pages.dev/tags/%E6%88%90%E6%9C%AC%E4%BC%98%E5%8C%96/index.xml" rel="self" type="application/rss+xml"/><item><title>从改良到重塑：解构 Prometheus 监控架构的三种哲学与选型真相</title><link>https://shengxu.pages.dev/posts/prometheus-monitoring-architecture-evolution/</link><pubDate>Sun, 04 Jan 2026 17:00:00 +0800</pubDate><guid>https://shengxu.pages.dev/posts/prometheus-monitoring-architecture-evolution/</guid><category domain="https://shengxu.pages.dev/categories/observability/">Observability</category><description>&lt;p&gt;回望过去几年在可观察性领域的摸爬滚打，尤其是围绕 Metrics 体系的建设，感觉就像是一场漫长的架构修行。从最开始守着单机 Prometheus 还要担心磁盘爆满，到后来引入 Thanos 试图做“无限存储”，再到如今用 Mimir 重构整个监控中枢，这些经历散落在记忆里，甚至有些细节已经开始模糊。&lt;/p&gt;</description></item></channel></rss>