These last few years I run an experiment, started with this post back in 2012.
I’ve included in my travis.ci build, on top of the usual build errors, also all of my pages that link to an external resource that no longer exists. The current build returns 64 errors (most of them are 404s).
- Twitter dev docs
- Microsoft documentation
- Yandex documentation
Some basic stats:
- Total links: 1503
- Errors: 64
- Percentage of errors: 3.108%
Not that high. Note that the “Total links” also include a lot of internal links.
One error was for a web archive page which seemed peculiar. The cause: robots.txt Now I’m just speculating, but it seems like the site that gave 404 was indexed in it, but was later removed due to a change in the robots.txt