Link Checking Code -- comments below.

links_check.py

      session = FuturesSession(max_workers=8)
      session.mount("http://", requests.adapters.HTTPAdapter(max_retries=5))

      msg = "%s: testing %d web and %d local links" % (nbname, len(web_links), len(local_links))
      print >> sys.stderr, msg

      # Testing web links
      while True:
        futures = {link: session.get(link, verify=False, stream=True,
                                     headers={'User-Agent': self.USER_AGENT}, timeout=10.0)
                   for link, count in web_links.items()
                   if count < self.TRIES}
        if not futures:
          break

        for url, future in futures.iteritems():
          r = None
          try:
            r = future.result(timeout=20)
          except requests.ConnectionError:
            web_links[url] += 1
          except TimeoutError:
            future.cancel()
            web_links[url] += 1
          else:
            if r.status_code >= 200 and r.status_code < 300: #2xx
              del web_links[url]
            else:
              web_links[url] += 1

      # some weird DNS error causes threads to
      # freak out and not die no matter what
      # so we kill any unstopped threads manually
      for thread in session.executor._threads:
        thread._Thread__stop()

      msg = "Web links failed {nb}: {links}".format(nb=nbname, links = ", ".join(web_links))
      assert web_links == {}, msg

This is messy, but is the worst part of the whole process.

1. Scraping the links is fairly straightforward (just search for all http/https URLs in your notebook JSON).
2. Once you do that, you can run this code to check all the web_links - set it up as a defaultdict(int) with the URLs as keys. So it'd look like this:

web_links = {"https://www.google.com": 0}

3. Last but not least, FuturesSession comes from requests-futures