Add/investigate a link checker for the k8s website

[celestehorgan]

This is a Feature Request

In order to provide more accurate, useful content and to enable more resilient translation workflows, we need to first ensure that links are accurate and unbroken. The best way to do this is through automation (a link checker).

What would you like to be added

Add https://github.com/wjdp/htmltest as a Make command so that individuals can test locally. Eventually, this could scale up to deeper GitHub/Netlify integration.

Comments

At the SIG-Docs quarterly planning, we decided that link checking was a necessary prerequisite to Improving intra-site linking for translation (#19526 or a similar effort). There's no way to validate how successful re-implementing links might be unless we can validate that our existing links are correct.

[celestehorgan]

Based on where #20606 stands at the moment, I'd like to make the following recommendations:

1. Clean up the WIP PR, add some documentation to the Contribute section/README.md on using it, and merge it.
2. Open a second PR which corrects the initial set of links found by the link checker. This way when someone chooses to use the link checker, they get actual errors.

---

Some things to note:

1. I don't think checking external links is feasible at this time – it takes too long (~15s for a round-trip ping to the external link's server)
2. I don't think we want this to fail netlify builds, though it's possible if we want it to.
3. I don't think we want to automate translation link checking, though it might be possible in the future – I suspect their script(s) will pick these up every version.

---

Some things to decide before we move forward:

1. What directory should we set as the scope of the linkchecker? We can scan all of the generated HTML (/public) but this would include translation links as well.
2. Are we comfortable with (adding a linkcheck-config.toml file? Hugo doesn't provide a build-time flag for this setting.
3. Are we comfortable with the initial setup

---

Some things we could potentially do as follow-up issues/PRs:

• Remove (one by one) content redirects from the _redirects file and correct them, cleaning up our redirects and making it easier for new contributors.
• Assess whether using a different link strategy (which would help translations out more) is feasible AND test to see whether anything gets broken in the process.