Hello! I am running Pyflink 1.15 (on AWS) and I am trying to annotate a live kinesis stream with extra attributes from a iceberg table using the table api. I have:

INSERT INTO kinesis_output
SELECT lookup.extra_attribute, stream.*
FROM kinesis_input stream
INNER JOIN iceberg_table lookup
ON stream.reference = lookup.reference

This executes fine, but according to my dashboard, the iceberg_table source is "FINISHED" and the output stream is empty. Is this approach viable, or have I misunderstood flinks capabilities? If I misunderstand, what would be the correct approach to annotate the stream?