Have question on retrying mechanism in case of error, precisely db connectivity error. Seems like its only possible when establishing replication connection through 'slot.max.reties' setting. Kafka-connect Introduced support for retry in https://cwiki.apache.org/confluence/display/KAFKA/KIP-298%3A+Error+Handling+in+Connect. In order for a task to be retried it must throw RetriableException. I'm killing database and debezium task is just stopped on poll:
Is this planned to be supported anytime soon or am I missing some other setting?
I got response that retrying is planned but without any dates set yet. I'm thinking on implementing this - have you thought already on any direction in which this can be made?
Basically from what I see in kafka-connect from plugin perspective to use retrying is to throw RetriableException instead of root exception. Debezium is a source plugin so I would add possibility to wrap exceptions thrown from poll() into RetriableException based on some configuration - does that make sense?
Possible look at https://github.com/confluentinc/kafka-connect-jdbc/blob/master/src/main/java/io/confluent/connect/jdbc/sink/JdbcSinkTask.java - possible showcase of reusing kafka-connect's retrying functionality.