When using the cdc.py tool to pipe data to kafka it unfortunately fails for a table with a lot of changes.
If I use the MariaDB table which has much fewer changes (lets say, our users):
/usr/bin/cdc.py -u cdc_user --p123456 -h 127.0.0.1 -P 4001 production.users | /usr/bin/cdc_kafka_producer.py --kafka-broker=172.30.0.190:9092 --kafka-topic=users
It works properly (and still pegs the CPU at 100%).
But when I switch to another table (items) with more data/changes, it pegs the CPU at 100%, and nothing is ever sent to the kafka topic.
The server running the script is on AWS (m4.large).