You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "Shuyi Chen (Jira)" <ji...@apache.org> on 2025/02/21 06:43:00 UTC

[jira] [Created] (FLINK-37366) Allow configurable retry for Kafka topic metadata fetch

Shuyi Chen created FLINK-37366:
----------------------------------

             Summary: Allow configurable retry for Kafka topic metadata fetch
                 Key: FLINK-37366
                 URL: https://issues.apache.org/jira/browse/FLINK-37366
             Project: Flink
          Issue Type: Improvement
          Components: Connectors / Kafka
            Reporter: Shuyi Chen


For high availability, we adopted a multi-primary Kafka cluster setup, so the data of a Kafka topic will be in multiple physical clusters. In case of a kafka cluster failure, Flink pipeline should  continue to run w/o failure. Currently, Flink pipeline will fail due to SubscriberUtils.getTopicMetadata() throwing RuntimeException if a kafka cluster fails, causing the pipeline keep restarting. We propose to add a configurable retry policy in SubscriberUtils.getTopicMetadata(), so we can configure flink Kafka connector to tolerate kafka failure for longer period of time w/o restarting.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)