[
https://issues.apache.org/jira/browse/FLINK-36817?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17903569#comment-17903569
]
Levani Kokhreidze commented on FLINK-36817:
-------------------------------------------
Hi [~davidradl]
Thanks for getting back to me.
The reason I have not added it to _KafkaPartitionDiscoverer_ and __
_KafkaConsumerThread_ is because they are based on Source V1 API which is
deprecated and I thought it wasn't worth it to add the new feature there.
Of course I can add it there too if it's needed.
> Give users ability to provide their own KafkaConsumer when using
> flink-connector-kafka
> --------------------------------------------------------------------------------------
>
> Key: FLINK-36817
> URL: https://issues.apache.org/jira/browse/FLINK-36817
> Project: Flink
> Issue Type: New Feature
> Components: Connectors / Kafka
> Reporter: Levani Kokhreidze
> Priority: Major
> Labels: pull-request-available
>
> In certain scenarios users of the `KafkaSource` in the flink-connector-kafka
> might want to provide their own KafkaConsumer. Right now this is not possible
> as consumer is created in the
> [KafkaPartitionSplitReader|https://github.com/apache/flink-connector-kafka/blob/main/flink-connector-kafka/src/main/java/org/apache/flink/connector/kafka/source/reader/KafkaPartitionSplitReader.java#L97]
> which makes customisation impossible.
> Proposal is to let users pass `KafkaConsumerFactory` when building the
> KafkaSource.
> {code:java}
> public interface KafkaConsumerFactory {
> KafkaConsumer<byte[], byte[]> get(Properties properties);
> }{code}
> Builder will have a default implementation which creates the KafkaConsumer
> the same as it is done now.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
