在Logstash中解析嵌套的JSON字符串

Question

我正在以json格式登录logstash,我的日志有以下字段,每个字段都是一个字符串,atts字段是字符串化的json(注意:atts每次子字段都不同)

这是一个例子:

{"name":"bob","last":"builder", "atts":"{\"a\":111, \"b\":222}"}

我想解析它是这样的:

   {
     "name" => "bob",
     "last" => "builder"
     "atss" => {
          "a" => 111,
          "b" => 222}
   }

这是我的配置:

input { stdin { } }  

filter {
  json {
    source => "message"
    target => "parsed"
  }
}
 output { stdout { codec => rubydebug  }}

好的,现在我明白了:

{
    "@timestamp" => 2017-04-05T12:19:04.090Z,
    "parsed" => {
        "atss" => "{\"a\":111, \"b\":222}",
        "name" => "bob",
        "last" => "the builder"
    },
        "@version" => "1",
          "host" => "0.0.0.0"
}

我怎样才能将atts字段解析为json所以我收到:

{
    "@timestamp" => 2017-04-05T12:19:04.090Z,
    "parsed" => {
        "atss" => 
           {"a" => 111,
            "b" => 222},
        "name" => "bob",
        "last" => "the builder"
    },
        "@version" => "1",
          "host" => "0.0.0.0"
}

Answer 1

感谢@Alcanzar这就是我所做的

input { 
  stdin { } 
}  

filter {
  json {
    source => "message"
    target => "message"
  }
  json {
    source => "[message][atts]"
    target => "[message][atts]"
  }

}
 output { stdout { codec => rubydebug  }}

Answer 2

有一个json过滤器。只需将您要解析的字段和您想要的目标传递给它。

就像是：

json {
  source => "[parsed][atss]"
  target => "[parsed][newfield]"
}

我不确定您是否可以将 atss 作为新领域。它可能会或可能不会起作用。如果没有，请使用mutate过滤器remove_field和rename_field。