CI: refcache update policy and cron-job proposal

[chalin]

@svrnm @cartermp @austinlparker: Severin brought up the topic of "burden on PR contributors" of having to update the refcache file. I'd like to propose the following.

• If a PR introduces "lots" of new external links, then the PR must/should include an updated refcache file.
• For other PRs, feel free to merge the PR whether the refcache is fresh or not -- freshness is checked via a separate GH action so it is possible to merge w/o the action passing.

We can then have, say, a cron job that periodically update the website's refcache file. WDYT?

Tasks

• Create update scripts Refcache refresh (final for 23Q2) + prune task updated options and features #2960
• Setup a cron job to run, say, once every 4 weeks and attempt to refresh, say, 500 entries

[cartermp]

I like it a lot.

[chalin]

I'd also like to propose the following tweak to the build script: it should fail (it doesn't currently), if there's an invalid link or if a new external link results in a 404 (independent of the refcache freshness). How does that sound?

[cartermp]

Yep, works for me too. Can always tweak later if we need to.

[svrnm]

+1 let's do that and improve from there :)

[chalin]

@trask - the semconv daily build job is working great! I'm wondering if you'd be willing to help with this issue by writing a similar script that creates or commits to a "Refcache refresh" PR on a daily basis. I think that I'm getting a sense of how to automate this process, and now have the base scripts ready to support that process. If you aren't available to write such a script, that's ok, but I figured that I'd ask. (Otherwise, I'll try my hand at it when I next have the time.)