AI爬虫挤崩Gentoo Bugzilla
推荐指数 50.0 NO. 012 · 2026.08.09
发布2026/08/08Score108Comments70
为什么值得看
Gentoo Linux的Bugzilla缺陷追踪系统因AI爬虫过度抓取导致服务过载而被迫关闭。这暴露了开源基础设施在AI训练数据采集团队面前的脆弱性，维护者缺乏有效反制手段。
编辑判断
这不是第一次开源基础设施被AI爬虫压垮，Cloudflare之前也报告过类似攻击量级增长。Gentoo作为小众发行版尚且如此，Apache、Mozilla等更大规模的Bugzilla实例面临的风险只会更高。
对AI从业者有两个直接启示：一是如果你在做训练数据采集团队，现在就该建立白名单机制和robots.txt合规审查，否则下一个被封的可能是你的IP段；二是开源维护者需要考虑引入更精细的速率限制和爬虫识别，而不是简单封禁——误伤正常开发者比爬虫本身更致命。
社区反馈
负面 67 条评论
核心争论：AI爬虫泛滥成灾，但责任主体模糊，小型开源项目缺乏有效防御手段
相关内容
going on a bit of a tangent here - some discussion there seems to be imply that AI bots mostly use IPv4s, which makes sense to me given that they are probably bots hosted by some cloud. Whereas IPv6 may be more organic traffic from e.g. mobile users (not in this case probably)... which made me think
This is weird, every VPS I've rented in the last 8 years had IPv6, and some didn't even had IPv4.
If a problematic bot is on any mainstream VPS it will just get shut down.