格式化JSON中的java日志以在ELK堆栈中使用

我试图以JSON格式输出我的日志，这样我就不需要在我的ELK堆栈中使用过滤器了。它似乎确实起作用了

这是我的

import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import org.json.simple.JSONObject; 

public class MyApp {

   static Logger logger = LoggerFactory.getLogger(MyApp.class);

   @SuppressWarnings("unchecked")
   public static void main(String[] args) {
      // TODO Auto-generated method stub
      JSONObject obj = new JSONObject();
      JSONObject obj2 = new JSONObject();


      obj.put("name","foo");
      obj.put("num",new Integer(100));
      obj.put("balance",new Double(1000.21));
      obj.put("is_vip",new Boolean(true));

      obj2.put("ob2name","bar");
      obj2.put("ob2num",new Integer(200));

      obj.put("names", obj2);


      logger.info("{}", obj);

   }

}

输出到my file.log

{"@timestamp":"2016-07-27T17:51:08.331+01:00","@version":1,"thread_name":"main","logger_name":"MyApp","level":"INFO","level_value":20000,"HOSTNAME":"gman","message":"{\"names\":{\"ob2num\":200,\"ob2name\":\"bar\"},\"balance\":1000.21,\"is_vip\":true,\"num\":100,\"name\":\"foo\"}"}

输出到日志存储控制台

{
     "@timestamp" => "2016-07-27T16:51:08.331Z",
       "@version" => 1,
    "thread_name" => "main",
    "logger_name" => "MyApp",
          "level" => "INFO",
    "level_value" => 20000,
       "HOSTNAME" => "gman",
        "message" => "{\"names\":{\"ob2num\":200,\"ob2name\":\"bar\"},\"balance\":1000.21,\"is_vip\":true,\"num\":100,\"name\":\"foo\"}",
           "host" => "gman",
           "path" => "C:\\apps\\dots\\logs\\file.log"
}

我的问题

---

消息似乎是以字符串形式输出的，我如何才能将其输出为JSON，以便Elasticsearch能够提取各个字段并对它们进行索引，以便它们可以进行搜索

在日志存储的conf文件中，添加json过滤器：

filter {
    json {
        source => message
    }
}

---

或者，使用地图并进行如下思考：

Map<String, Object> map = new HashMap<>();
map.put("system", "fedora");
logger.info(Markers.appendEntries(map), "Hello world");

（省略默认字段）

{
  "message" : "Hello world",
  "system" : "fedora"
}