Ensures that a url's healthcheck is a success before marking it alive in the pool
Ensures that the version of the connected Elasticsearch is retrieved successfully before allowing the connection to be marked as alive
Eliminates a crash that can occur when HTTP connection can be established and the connected resource isn't Elasticsearch (e.g., Elastic Cloud returning an HTTP 404 Not Found)

* logging: ensure we log sanitized url

While our logging framework is known to send `Object#to_s` to the
objects we pass it, we should not rely on that and shoukd instead
consistently sanitize our url when logging.

* pool: require SUCCESSFUL http response for connection

The Manticore Adapter we provide counts an HTTP 404 Not Found to be
non-exceptional, which causes our healthcheck connection reestablishing
to succeed when the server it connects to reports that the endpoint
cannot be found.

Co-authored-by: Ry Biesemeyer <ry.biesemeyer@elastic.co>
Co-authored-by: Mashhur <99575341+mashhurs@users.noreply.github.com>