collector: pass Context

[SaveTheRbtz]

This diff adds an ability to pass custom Context to the collector. This is inline with some other exporter's like the mysqld_exporter.

[SuperQ]

Ahh, yes, it would be useful use the X-Prometheus-Scrape-Timeout-Seconds header.

I'm not sure how useful it is to stub this out, without implementing any use of the context is. Are you planning to add more to this PR?

[SaveTheRbtz]

We use collector directly as a library so we pass our own context therefore I marked the main.go context as TODO. I you want I can add support for X-Prometheus-Scrape-Timeout-Seconds in this diff, or a subsequent one.

[SuperQ]

Yes, with normal use, no context is necessary, as the scrape drops the HTTP connection on timeout, which cancels the walk. I was wrong, we don't need to explicitly handle the X-Prometheus-Scrape-Timeout-Seconds here. It's automatic.

[SaveTheRbtz]

Ok. I've added context propagation from the original HTTP request to the underlying collector. This should ensure cancelation in following scenarios:

For incoming server requests, the context is canceled when the client's connection closes, the request is canceled (with HTTP/2), or when the ServeHTTP method returns.

[SuperQ]

Thanks, it looks like I screwed up the context here. This puts it back to where it was.

[SuperQ]

LGTM.