想必昨晚上网尤其是访问一些国际化大站的朋友会比较闹心。为什么这么说呢？因为很多网站都没法正常访问，当然也包括我们的hive.blog。
是hive.blog出了什么故障嘛？一些小伙伴私聊我询问这个问题，还有小伙伴在微信群发出疑问！答案是：虽然hive.blog无法访问，但是这并不是hive.blog的问题，而是Cloudflare的故障。
听起来有点绕，是不是？别急，听我慢慢道来。
一般来讲，一个网站要在互联网上能被访问，需要有个运行网站的服务器（或者虚拟主机空间），一个绑定到这个服务器的IP地址，还要有一个域名指向(DNS)这个IP。
比如我的cutehive.com网站就是如此，在主机服务商处创建了VPS，服务商同时分配了172.234.182.66这个IPv4地址，我使用域名解析商的DNS服务，将域名cutehive.com解析到这个IP。喏，现在你就可以打开域名访问我的网站了。
很多小网站几乎都是这个模式，成本低廉，部署简单高效，也不用考虑太多的问题，因为原本就没多少访问，哈哈哈。
但是对一些大网站、甚至超大网站而言，事情就变得略微复杂一些了。为了让网站能承受每分钟成千上万乃至上亿次的访问，为了让全球各个地区都能迅速的响应，就要引入负载均衡以及CDN等机制。
同时，直接暴露网站的IP也会带来诸多风险，比如被攻击者用一些工具进行漏洞扫描，又比如被DDOS攻击，还要面对一些疯狂的爬虫等等。
Cloudflare就能让你一劳永逸地解决上述烦恼，帮你做域名解析，帮你隐藏源站 IP（保护真实服务器），帮你做CDN 加速（让网站更快），帮你防御 DDoS 攻击，还可以帮你页面缓存、压缩、优化等等。
（HIVE的DNS设置：来源)
你想到的想不到的它都能帮你做，而且在这方面它们做的确实也不错。所以，全球最牛叉的10000个网站中，有4332个(43.32%)使用了Cloudflare的服务。
In total, 24,030,148 active websites use Cloudflare worldwide. Ranked by traffic4,332 out of 10,000 most popular websites globally use Cloudflare.
从昨天出故障的网站来看，不止hive.blog，x.com，chatgpt.com，truthsocial.com等，都或多或少用了Cloudflare服务。
这些站和hive.blog一样，遇到Cloudflare出现严重故障，就无法正常访问了。
当然，这都不是站点本身的问题，原因就是我封面图中左下侧的提示(是Cloudflare网络的故障）：
There is an internal server error on Cloudflare's network.
这些大的服务商做的很赞的一点就是，当故障发生时，故障的处理进展会在网站上即时更新。看到工程师们不懈努力地排查和解决故障，给人一种安心的感觉。
上图是昨晚Cloudflare网站上的状态更新。到我睡觉前，工程师们已经修了三个多小时啦（睡觉前忘记截图啦）。
好在早晨起来，看到一切都OK了，找到了关于这次全球DOWN机的最新报告：
上边是最新部分的截图，想看完整进度的请点击：
好消息是，Cloudflare终于宣布：This incident has been resolved.
坏消息是，从故障发生Nov 18, 2025 - 11:48 UTC 到彻底解决Nov 18, 2025 - 17:44 UTC ，整整耗时6个小时。考虑到全球大站对Cloudflare的依赖性，这次Down机造成的损失将会难以估量。
(图源 ：pixabay)
另外一个好消息是，尽管我们常用的HIVE前端hive.blog在这期间受到影响没法用来发帖点赞等，但HIVE有好些个前端和DAPP，有些不依赖于Cloudflare就没有受到影响。
还有就是基于HIVE的前端和DAPP和HIVE区块链是两码事，HIVE区块链运行在诸多分布式节点之上，完全不受这次Cloudflare DOWN机的任何影响。
目前HIVE TOP20见证人节点已经全部部署运行了1.28.3以上版本的hived，其它备选（分时）节点绝大部分也都更新到hived的最新版本。
我也备好咖啡，坐等今晚的硬分叉哦（北京时间19号21时）。不清楚具体时间的小伙伴，可以参考https://hive.io/首页的倒计时哦。
It was a rough morning for a lot of us! I reached out to some of my colleagues and I sent messages out to my end users letting them know things would be "flaky" for a bit. I'm actually surprised it took them as long as it did to get things fixed.
A global outage like this is really quite terrible.
I saw all kinds of complaints from friends in many WeChat groups. Not only were some major websites inaccessible, but many people’s own websites and scripts also stopped working. The actual losses are hard to estimate.
Of course, occasional issues are unavoidable, but a failure with such a massive impact really shouldn’t happen to a company of this scale. And taking more than six hours to investigate and fix the problem was indeed rather inefficient.
Yeah, it was pretty out of hand. It sounded like they were doing some routine maintenance and then things just kind of went sideways on them. It's crazy just how much relies on them working right. I work in public education and I would guess half our services that we don't host locally were inaccessible.
Some services, such as login verification, are deployed on Cloudflare.
Then there are other websites that aren’t hosted on Cloudflare but rely on those services, so they also became unusable — like ChatGPT.
So the impact was really huge.
Ah yes, I can only imagine. I actually had vendors sending emails out touting that they don't use Cloudflare so they stayed stable all morning. Not a bad marketing opportunity for them!
Haha, just like my cutehive site —
there’s absolutely no need to use Cloudflare. 😀
Faced out with cloudflare issue yesterday when published new post on hive. First foughts were that something happened with my vless server or hive has been blocked by provider and local authorities, but then read about cloudflare issue and understood that here we can just wait.
Haha, many people thought the problem was caused by some part of HIVE.
❤️谢谢O神大佬分享☕☕，了解了不少知识👍👍👍
一起学习进步
这是一份清晰而详细的说明，表明我们在互联网访问方面过于依赖后端服务。希望 Cloudflare 能尽快解决这个问题，恢复所有受影响网站的访问……
This a clear and detailed explanation of this issue. It shows our over reliance on backend stuff for internet access. Hopefully cloudflare resolves this quickly to restore access to all websites affected...
谢谢O哥让我们知道更多这方面知识！😃😃
一起学习进步
又是涨知识的一天～
一起学习进步
我的algo grade昨晚突然連不上, 到現在還未解決, 應該也是跟這有關🥹
嗯嗯其实对我来说还好吧！其他平台也有这个情况出现，习惯了。毕竟国内不比国外，啥啥都被jian kong 着。不过现在能用就行！谢谢哥的分享学术文章🙏
我这几天也时不时出现打不开的现象，坐等硬分叉
发生了这么多大事啊！😱
涨知识了，谢谢O哥分享