Hi All, I have a mysql to elasticsearch data indexing setup with a mysql debezium connector with 4 consumer threads, a java application to transform the messages read by the debezium connector and a ES sink connector with 5 topics haveing the following configuration “tasks.max”: “4”,
“flush.timeout.ms”: “200000”,
“read.timeout.ms”: “20000”,
“linger.ms”: “20”,
“batch.size”: “1000”
I have set the Kafka producer batch size to 100000 bytes ( Just trying out different sizes to see the difference in performance)
and linger.ms to 20
For Kafka Consumer , I have set the MAX_POLL_RECORDS_CONFIG to 100 and MAX_POLL_INTERVAL_MS_CONFIG to maxValue
After the pipeline runs for a few days the indexing speed slows down . I am not completely sure what is the reason as the logs seem fine for kafka and kafka connect. Any help/suggestions are appreciated . Thanks In advance