Pass retry_on_conflict parameter for fluentd

I had configured fluentd to listen to messages from rabbitmq and persist into elasticsearch 7x. I see my records are gettting flushed in a bulk request the request body is

{"update":{"_index":"cat","_type":"fluentd","_id":"1127429471072048942891864682698024"}}
{"doc":{"trace_id":"26a166fb75ec36f8"},"doc_as_upsert":true}

how can I pass the retry_on_conflict parameter by configuring it in fluentdcp template.
There is a 409 exception I am getting and if I can pass this parameter then it will retry