On 2026-03-18 08:08:57 +0000 (+0000), Eugen Block wrote:
I'm not sure how long it has been like that, but when I tried to access the documentation yesterday, it took forever to load. And it is the same today, it takes several minutes to load any site, for example:
https://docs.openstack.org/cinder/2025.2/admin/
I tried it from different networks, didn't make a difference. Is that a known issue?
OpenStack's documentation (along with many other sites) is hosted in the volunteer-run OpenDev Collaboratory, which for the past week has been weathering a massive attack from the latest iteration of the "AI Goldrush." So far we've scaled up the virtual machine resources serving that content by a factor of 600% and implemented adaptive web application firewall rules to block crawlers we can identify, but this particular swarm is something we've never seen before. Hundreds of thousands of IP addresses in networks around the World (probably backdoored mobile devices judging from the networks where they originate) are masquerading as normal web browsers while requesting random permutations of URLs, most of which end up at the blanket redirects OpenStack's docs maintain for backward compatibility over ~16 years of URL changes. Sadly, the OpenStack community lost all of its documentation maintainers about a decade ago, so there's currently no one to help refactor and rework all of its htaccess-based redirect rules and related management tooling in order to help relieve that aspect of the problem. The OpenDev sysadmins are continuing to tune the performance for these servers and are working on other possible mitigations as well, some of which may unfortunately make the documentation less accessible overall as a tradeoff for keeping it accessible at all. We're still going to do our best to avoid compromising our users' privacy, by sticking to self-hosted solutions to the problem for as long as that remains viable. -- Jeremy Stanley, on behalf of the TaCT SIG and OpenDev Sysadmins