If you run multiple Kafka connector instances on the same topics or partitions, what is likely to happen?

• A. Multiple copies of the same row may be inserted
• B. No duplicates will occur
• C. Data will be lost
• D. Processing will be faster

Answer: (A)

When you run several connector instances against the same topics or partitions, you introduce parallel processes that may process the same data more than once. Each connector can independently read and write, so the same record can be ingested into the destination multiple times if there isn’t a global coordination or deduplication mechanism. Kafka’s default delivery is at-least-once, and without idempotent writes or exactly-once semantics across the connectors, duplicates are a real risk. To reduce this, isolate connectors by using separate topics or partitions, or implement deduplication in the sink (or enable transactional/exactly-once processing if supported).