Kafka Connect MySQL to Redshift - Dates Cause Error

Question

I am getting this error when attempting to stream an RDS MySQL table into Redshift: Error converting data, invalid type for parameter

The problem field is a DATETIME in MySQL and timestamp without time zone in Redshift (same happens for timestamp with time zone). Note: pipeline was working fine until I populated the date field.

We are using Debezium as the Kafka Connect source for getting data from RDS into Kafka. And the JDBC sink connector with Redshift JDBC driver for the sink.

Also... I am able to get the data flowing if I make the Redshift field a varchar or a bigint. When I do this, I see that the data is coming across as a unix epoch integer in ms. But we'd really like a timestamp!

Error message in context:

2018-10-18 22:48:32,972 DEBUG  ||  INSERT sql: INSERT INTO "funschema"."test_table"("user_id","subscription_code","source","receipt","starts_on") VALUES(?,?,?,?,?)   [io.confluent.connect.jdbc.sink.BufferedRecords]
2018-10-18 22:48:32,987 WARN   ||  Write of 28 records failed, remainingRetries=7   [io.confluent.connect.jdbc.sink.JdbcSinkTask]
java.sql.BatchUpdateException: [Amazon][JDBC](10120) Error converting data, invalid type for parameter: 5.
    at com.amazon.jdbc.common.SStatement.createBatchUpdateException(Unknown Source)
    at com.amazon.jdbc.common.SStatement.access$100(Unknown Source)
    at com.amazon.jdbc.common.SStatement$BatchExecutionContext.createBatchUpdateException(Unknown Source)
    at com.amazon.jdbc.common.SStatement$BatchExecutionContext.createResults(Unknown Source)
    at com.amazon.jdbc.common.SStatement$BatchExecutionContext.doProcess(Unknown Source)
    at com.amazon.jdbc.common.SStatement$BatchExecutionContext.processInt(Unknown Source)
    at com.amazon.jdbc.common.SStatement.processBatchResults(Unknown Source)
    at com.amazon.jdbc.common.SPreparedStatement.executeBatch(Unknown Source)
    at io.confluent.connect.jdbc.sink.BufferedRecords.flush(BufferedRecords.java:138)
    at io.confluent.connect.jdbc.sink.JdbcDbWriter.write(JdbcDbWriter.java:66)
    at io.confluent.connect.jdbc.sink.JdbcSinkTask.put(JdbcSinkTask.java:75)

Thanks,

Tom