Currently

Cljdoc offers the ability to rebuild docs via an almost hidden rebuild link.

Why did we choose to make it almost invisible? I'm pretty sure it is:

• because we did not to draw attention to it
• because it is not relevant/appropriate for most cljdoc users
• and because it invokes a significant amount of work

So...

Rate limit this heavy work?

@corasaurus-hex suggested maybe we should rate limit the rebuild feature.
And I agreed this is a good idea.

We don't know of anybody abusing rebuild yet, but I think some sort of protection makes sense here.

We could start very conservative and only allow one rebuild to be active at a time with no queueing support. And then adapt based on user feedback.

Move it to where it is more relevant?

I also think we should move the rebuild link to the build page.
I expect most cljdoc users will not navigate to the build page.

We could replace the "rebuild" link with a visible "build" link which would bring you build page (which is currently only linked for failed builds). Leaving the link in the same location would mean folks used to hitting rebuild would not be lost.

A visible "rebuild" link would be added to the build page.
It might pop up an "are you sure?".
It would provide info about our rate limiting.

Thoughts?

I think the above is a good start, but if you have ideas, please do share.