tilecache: reject request without user-agent (try2)

[chef.git] / cookbooks / tilecache / templates / default / nginx_tile.conf.erb

diff --git a/cookbooks/tilecache/templates/default/nginx_tile.conf.erb b/cookbooks/tilecache/templates/default/nginx_tile.conf.erb
index 95d7b8b8c062f559c22d3de12534da4740e90094..3932f54b7b0462a9a4d3dac16ae8f3037488be89 100644 (file)
--- a/cookbooks/tilecache/templates/default/nginx_tile.conf.erb
+++ b/cookbooks/tilecache/templates/default/nginx_tile.conf.erb
@@ -1,9 +1,31 @@
 # DO NOT EDIT - This file is being maintained by Chef
 
 upstream tile_cache_backend {
-    server 127.0.0.1;
+  server 127.0.0.1;
+
+  # Add the other caches to relieve pressure if local squid failing
+  # Balancer: round-robin
+<% @caches.each do |cache| -%>
+<% if cache[:hostname] != node[:hostname] -%>
+<% if node[:tilecache][:tile_siblings].include? cache[:fqdn] -%>
+<% cache.ipaddresses(:family => :inet, :role => :external).sort.each do |address| -%>
+  server <%= address %> backup; # Server <%= cache[:hostname] %>
+<% end -%>
+<% end -%>
+<% end -%>
+<% end -%>
 
-    keepalive 32;
+  keepalive 256;
+}
+
+# Geo Map of tile caches
+geo $tile_cache {
+  default 0;
+<% @caches.each do |cache| -%>
+<% cache.ipaddresses(:family => :inet, :role => :external).sort.each do |address| -%>
+  <%= address %> 1; # <%= cache[:hostname] %>
+<% end -%>
+<% end -%>
 }
 
 # Rates table based on current cookie value
@@ -17,9 +39,36 @@ map $cookie_qos_token $cookie_qos_token_set {
 }
 
 map $http_user_agent $approved_scraper {
-  default '';                       # Not approved
-  '~^JOSM\/' 'JOSM';
-  '~^Mozilla\/5\.0\ QGIS\/' 'QGIS';
+  default                   0; # Not approved
+  '~^JOSM\/'                1; # JOSM
+  '~^Mozilla\/5\.0\ QGIS\/' 1; # QGIS
+}
+
+map $http_user_agent $denied_scraper {
+  default                0; # Not denied
+  ''                     1; # No User-Agent Set
+  '~^Python\-urllib\/'   1; # Library Default
+  '~^python\-requests\/' 1; # Library Default
+  '~^node\-fetch\/'      1; # Library Default
+  '~^R$'                 1; # Library Default
+  '~^Java\/'             1; # Library Default
+  '~^tiles$'             1; # Library Default
+  '~^Dalvik\/'           1; # Library Default
+  '~^runtastic'          1; # App
+  'Mozilla/4.0'          1; # Fake
+  'Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1)' 1;  # Fake
+}
+
+map $http_referer $denied_referer {
+  default                          0; # Not denied
+  'http://www.openstreetmap.org/'  1; # Faked
+  'http://www.openstreetmap.org'   1; # Faked
+  'http://openstreetmap.org/'      1; # Faked
+  'http://openstreetmap.org'       1; # Faked
+  'http://www.osm.org/'            1; # Faked
+  'http://www.osm.org'             1; # Faked
+  'http://osm.org/'                1; # Faked
+  'http://osm.org'                 1; # Faked
 }
 
 # Limit Cache-Control header to only approved User-Agents
@@ -79,7 +128,16 @@ server {
 
       # Allow Higher Traffic Rate from Approved User-Agents which do not support cookies (uses nginx Map)
       if ($approved_scraper) {
-        set $limit_rate 32768;
+        set $limit_rate 65536;
+      }
+
+      if ($denied_scraper) {
+        set $limit_rate 512;
+        return 429;
+      }
+      if ($denied_referer) {
+        set $limit_rate 512;
+        return 418;
       }
 
       # Strip any ?query parameters from urls