Improve link checking

[zacharykeeping]

This implements a few changes to the link scanning code to improve accuracy and cut down on the amount of false positives:

1. Switches to GET by default as we can confirm that we're not downloading much additional data as we're not handling the response bodies of requests. This also includes code to cancel and close any connections after we've recorded the response status data to be safe.
2. Updates headers on requests to better match that of a web browser to ensure results are more accurate.
3. Updates the timeouts in the http client to prevent connections ending too quickly.
4. Cuts down on the amount of "0 - No Response" errors as a lot of these seem to be simply connections closing and not actual errors. This now specifically checks for DNS errors (ie if a host no longer exists) and records these as Host Errors while filtering out other errors.
5. Ignores 429 Too Many Request errors since we're not able to get accurate results so we cannot definitively count them as broken. We'll have to investigate ways around this going forward (time retries, etc.).
6. Updates Go to the latest version.

#674 #732

[tombui99]

LGTM