Hello,
I have a table and a pipeline that inserts data from Kafka.
CREATE OR REPLACE PIPELINE player_audit
AS LOAD DATA KAFKA 'cluster/topic'
BATCH_INTERVAL 30000
SKIP DUPLICATE KEY ERRORS
INTO TABLE player_audit
FORMAT JSON
(
@v_data <- data,
)
SET
... fields set using JSON EXTRACT ...
;
Sometimes this pipeline misses events.
I stopped the pipeline, checked and saw 3 events in my kafka topic, but using test pipeline
I got only 1 event in the pipeline.
After running ALTER PIPELINE player_audit SET OFFSETS LATEST;
it suddenly worked OK.
It is important to note that this is development cluster - not prod - so it is not that powerful.
memsql 7.0.9
2 nodes that are not working that hard.
Any ideas why this might have happened?