How to perform aggregation in fluentd on nested data

Question

I am sending below Json message on Fluentd input.

{"name":"my_counter","help":"A counter metric","type":0,"metric":\[{"label":\[{"name":"pod","value":"test"}\],"counter ":{"value ":6}}\]}

And I wanted to perform the aggregation on the above data.

<source>
  @type dummy
  dummy {"name":"my_counter","help":"A counter metric","type":0,"metric":[{"label":[{"name":"pod","value":"test"}],"counter ":{"value":6}}]}
  tag test
  rate 1
</source>
<filter test>
  @type aggregate
  intervals 5s
  keep_interval 1s
  group_fields counter
  aggregate_fields value
  aggregations sum
</filter>
<match test>
  @type stdout
</match>

Aggregation of the sameple data coming at the rate of 1s for 5s interval should results into

{"name":"my_counter","help":"A counter metric","type":0,"metric":[{"label":[{"name":"pod","value":"test"}],"counter ":{"value":30}}]}

How can we perform aggregation in fluentd on nested data.