您可以使用带有key_name、reserve_data和remove_key_name字段的parser过滤器插件。
示例：

<filter **>
  @type parser
  key_name field3
  reserve_data true
  remove_key_name_field true
  <parse>
    @type json
  </parse>
</filter>

以下是使JSON有效后完整的工作示例，即：

{"field1":"value1","field2":"value2","field3":"{\"nested_field1\":\"value1\",\"nested_field2\":\"value2\",\"nested_field3\":\"value3\"}"}

流畅-展平-json.配置

<source>
  @type forward
</source>

<filter **>
  @type parser
  key_name field3
  reserve_data true
  remove_key_name_field true
  <parse>
    @type json
  </parse>
</filter>

<match **>
  @type stdout
</match>

运行fluentd：

fluentd -c ./fluent-flatten-json.conf

从另一个终端运行fluent-cat，输入JSON：

fluent-cat test <<< '{"field1":"value1","field2":"value2","field3":"{\"nested_field1\":\"value1\",\"nested_field2\":\"value2\",\"nested_field3\":\"value3\"}"}'

fluentd日志中的输出：

{"field1":"value1","field2":"value2","nested_field1":"value1","nested_field2":"value2","nested_field3":"value3"}

格式化输出：

{
  "field1": "value1",
  "field2": "value2",
  "nested_field1": "value1",
  "nested_field2": "value2",
  "nested_field3": "value3"
}