<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>容灾 on 硅基 AGI · 智能体学习与测评</title><link>https://guijiagi.com/tags/%E5%AE%B9%E7%81%BE/</link><description>Recent content in 容灾 on 硅基 AGI · 智能体学习与测评</description><generator>Hugo</generator><language>zh-cn</language><copyright>本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权</copyright><lastBuildDate>Mon, 13 Jul 2026 05:30:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E5%AE%B9%E7%81%BE/index.xml" rel="self" type="application/rss+xml"/><item><title>AI Agent的容灾与高可用设计：从理论到落地</title><link>https://guijiagi.com/posts/article-76/</link><pubDate>Mon, 13 Jul 2026 05:30:00 +0800</pubDate><guid>https://guijiagi.com/posts/article-76/</guid><description>深入探讨AI Agent系统的高可用架构设计，涵盖多活部署、故障转移、状态恢复与降级策略</description></item><item><title>Agent 降级策略：当 LLM 不可用时的容灾方案</title><link>https://guijiagi.com/posts/agent-degradation-strategy/</link><pubDate>Sun, 28 Jun 2026 10:35:00 +0800</pubDate><guid>https://guijiagi.com/posts/agent-degradation-strategy/</guid><description>构建Agent系统的多级容灾体系，从模型降级到规则回退，确保LLM不可用时服务不中断</description></item><item><title>LLM 服务容灾设计：多模型/多区域/降级策略</title><link>https://guijiagi.com/posts/llm-disaster-recovery/</link><pubDate>Thu, 25 Jun 2026 10:00:00 +0800</pubDate><guid>https://guijiagi.com/posts/llm-disaster-recovery/</guid><description>系统设计 LLM 服务的高可用架构，涵盖单点风险分析、多模型冗余、多区域部署与优雅降级策略</description></item></channel></rss>