小编exc*_*id3的帖子

我们有一台服务器正在提供一个html文件.

现在服务器有2个CPU和2GB内存.从blitz.io开始,我们每分钟可以获得大约12k个连接,并且在60秒内有200个超时,每秒250个并发连接.

worker_processes  2;

events {
 worker_connections 1024;
}

如果我增加超时,响应时间开始爬升超过一秒.

我还能做些什么来挤出更多汁液？

我正在尝试使用Basecamp Classic API(http://developer.37signals.com/basecamp/comments.shtml).当前的basecamp-wrapper版本给了我适合,其中一个原因是因为json响应包括分页输出,而xml响应包括分页输出.这是一个简单的解决方案,但问题是网址结构不规范.

API指定了一些类似的东西,这使我相信它只是分离出元素路径和集合路径.

Get recent comments (for a commentable resource)
GET /#{resource}/#{resource_id}/comments.xml

Update comment
PUT /comments/#{id}.xml

Run Code Online (Sandbox Code Playgroud)

我已经做了几次尝试,并没有真正成功.尝试处理这样的注释充其量只是hacky,并且实际上并不起作用,因为element_path与collection_path不同.

class Resource < ActiveResource::Base
  self.site = "https://XXXX.basecamphq.com"
  self.user = "XXXX"
  self.password = "X" # This is just X according to the API, I have also read nil works
  self.format = :xml # json responses include pagination crap

  # Override element path so it isn't nested
  class << self
    def element_path(id, prefix_options={}, query_options={})
      prefix_options, query_options = split_options(prefix_options) if query_options.nil?
      "#{collection_name}/#{URI.parser.escape id.to_s}.#{format.extension}#{query_string(query_options)}" …

我正在使用像这样的gmaps4rails:

  <%= gmaps("map_options" => { "scrollwheel" => false, "zoom" => 17, "auto_zoom" => false },"markers" => { "data" => event.to_gmaps4rails }) %>

这产生:

<script type="text/javascript" charset="utf-8">
Gmaps.map = new Gmaps4RailsGoogle();
Gmaps.load_map = function() {
Gmaps.map.map_options.scrollwheel = false;
Gmaps.map.map_options.zoom = 17;
Gmaps.map.map_options.auto_zoom = false;
Gmaps.map.initialize();
Gmaps.map.markers = [{"lat":38.6419,"lng":-90.2646}];
Gmaps.map.create_markers();
Gmaps.map.adjustMapToBounds();
Gmaps.map.callback();
};
window.onload = function() { Gmaps.loadMaps(); };
</script>

我忽略了这个scrollwheel = false选项我做错了什么？