Gentoo’s Bugzilla went dark because AI scrapers made it unusable. A Gentoo developer pulled the plug rather than fight distributed LLM crawlers.
The specifics, from Michał Górny’s post:
- Thousands of IPv4 addresses hitting Bugzilla with no obvious pattern to block
- No rate-limiting fix available — the address spread made IP bans or throttling impractical
- Górny isn’t a sysadmin — he’s a developer trying to get actual work done, not mitigate scraping infrastructure
- Bugzilla was already unusable before the takedown — the shutdown just made the failure official
The scrapers weren’t targeting Gentoo’s bug database specifically. They were vacuuming up public web content for LLM training data. Bugzilla instances are easy marks: public, structured, text-heavy, rarely hardened against automated access at scale.
This is the operational tax that AI tooling imposes on open-source infrastructure. The people paying it aren’t the ones building AI tools. Volunteer maintainers don’t have CDN budgets or SRE teams. Górny’s frustration is the whole story: “I’m just trying to get some useful job done. I’m not supposed to have to be dealing with this.”
NOTE
The HN thread hit 111 points and 75 comments, with commenters sharing similar scraper overload stories across other open-source projects.
No fix exists from the maintainer’s side. Górny took Bugzilla offline because the alternative was spending his time fighting crawlers instead of doing his actual job — and the next maintainer to face this choice won’t have better options.