支持日志解析，解析出来的字段可以并转换为数组

[zjf88865]

Describe the feature

举例 日志内容为： 01-Dec-2021 03:13:58.298 INFO shenzhen [main] Starting service [Catalina] time_zone:G8 country:China { "time": "01-Dec-2021 03:13:58.298", "fields": { "logconfig": "tomcat", "namespace": "test1", "nodename": "kind-control-plane", "podname": "tomcat-85c84988d8-frs4n", "containername": "tomcat", "desc" :"这里是无用的信息，哈哈哈啊哈哈哈哈哈哈" }, "level": "INFO", "server_area": "shenzhen" "message": "[main] Starting service [Catalina] time_zone:G8 country:China" "keys":["time_zone","country",${server_area},${fields.namespace}], "values":["G8","China","shenzhen","test1"] } keys和values可以随意命名,keys数组内容出现和下标和values出现的小标对应，例如：time_zone和G8,还要支持从现有输出格式里找到内容进行二次解析，上面例子就是我从message里解析出来的,${}可以从当前格式里取数据，例如${fields.namespace} 取到容器命名空间,${server_area}取到服务器部署区域，支持map取值，数组取值，多层级取值(a.b.c[0].d),解析如果发生错误，不能影响keys和values对应长度，例如["time_zone","country",${server_area},${fields.namespace}]，我命名空间拿不到异常了，应该为null,而不是只有3个，如果可以最好解析完成后再支持后续解析去掉无用信息，例如我要去掉fields.desc { "time": "01-Dec-2021 03:13:58.298", "fields": { "logconfig": "tomcat", "namespace": "test1", "nodename": "kind-control-plane", "podname": "tomcat-85c84988d8-frs4n", "containername": "tomcat" #这行数据被去掉了 }, "level": "INFO", "server_area": "shenzhen" "message": "[main] Starting service [Catalina] time_zone:G8 country:China" "keys":["time_zone","country",${server_area},${fields.namespace}], "values":["G8","China","shenzhen","test1"] }