remote-blackbox: Do not follow HTTP redirects

There are a couple of websites we scrape that simply redirect to another
name (e.g. _pyrocufflink.net_ → _dustin.hatch.name_, _tabitha.biz_ →
_hatchlearningcenter.org_).  For these, we want to track the
availability of the first step, not the last, especially with regard to
their certificate lifetimes.

group_vars/needproxy.yml

@@ -1,4 +1,4 @@
 http_proxy: http://proxy.pyrocufflink.blue:3128
 https_proxy: '{{ http_proxy }}'
 all_proxy: '{{ http_proxy }}'
-no_proxy: localhost,pyrocufflink.blue,*.pyrocufflink.blue,127.0.0.1,172.30.0.*,172.30.0.0/24
+no_proxy: localhost,pyrocufflink.blue,logs.pyrocufflink.blue,*.pyrocufflink.blue,127.0.0.1,172.30.0.*,172.30.0.0/24

group_vars/remote-blackbox.yml

@@ -6,6 +6,15 @@ blackbox_modules:
         Accept-Charset: utf-8
         Accept-Language: en-US
       method: GET
+      follow_redirects: false
+      valid_status_codes:
+      - 200
+      - 301
+      - 302
+      - 303
+      - 304
+      - 307
+      - 308
     prober: http
     timeout: 5s