A source connector moves data from an external system into Kafka. A sink connector moves Kafka records toward another system. Connector tasks divide supported work, but tasks.max is a ceiling rather than a guarantee that every connector creates that many useful tasks. Source offsets track progress in the source system and differ from Kafka partition offsets. These exercises stage a fictional source row, run the source connector and inspect the appended record and saved source position. Local controls replace the Connect REST API and connector plugins, so no external system is contacted. A staged row alone is not an ingested event. Running the task is the operation that advances the pipeline.
Inspect the topics, partition logs and group offsets before changing the cluster.