Http Sink Connector retry messages when the service has a long-time response

[NM-Narut]

I have to created sink connector with configuration like this
{
"connector.class": "io.aiven.kafka.connect.http.HttpSinkConnector",
"http.authorization.type": "none",
"tasks.max": "3",
"name": "{{connector-name}}",
"http.url": "{{service-endpoint-url}}",
"auto.commit.interval.ms": "15000",
"heartbeat.interval.ms": "15000",
"value.converter": "org.apache.kafka.connect.storage.StringConverter",
"retry.backoff.ms": "30000",
"http.ssl.trust.all.certs": "true",
"topics.regex": "{{topic-name}}",
"max.poll.interval.ms": "3600000"
}

I reviewed the service response log and observed that the duration ranged between 5 and 10 seconds, then I counted the number of requests and I found the number of requests was more than the number of messages. Assume that if I have 500 messages in Kafka, the number of requests is more than 500 requests and I find some messages will be sent duplicate.

Which config will fix this issue ?

Thank you

[chatchaisi]

I already have this issue when a consumer group is rebalancing.
API Response time around 1-2 sec.

I don't know how to fix this. If you have solution to fix this. Please reply how to resolve it to me.

[sp1rs]

Probably because you the max.poll.interval.ms will not have any effect on the kafka connector. the correct way to pass consumer configurations is prefixing it with consumer., so it should be consumer.max.poll.interval.ms.