**At-most-once**: Producer fires and forgets. Messages may be lost. Use for metrics where loss is acceptable.
**At-least-once**: Producer retries until ack; consumer may see duplicates. Simpler; use with idempotent consumer.
**Exactly-once**: Transactional producer + transactional consumer (Kafka 0.11+). Producer commits atomically; consumer reads-processes-commits in transaction. No duplicates, no loss.
**Why It Matters**: Financial transactions, billing, inventory need exactly-once....
The complete answer continues with detailed implementation patterns, architectural trade-offs, and production-grade considerations. It covers performance optimization strategies, common pitfalls to avoid, and real-world examples from companies like Fragma Data Systems. The answer also includes follow-up discussion points that interviewers commonly explore.
Continue Reading the Full Answer
Unlock the complete expert answer with code examples, trade-offs, and pro tips - plus 1,863+ more.
Or upgrade to Platform Pro - $39
Engineers who used these answers got offers at
AmazonDatabricksSnowflakeGoogleMeta
According to DataEngPrep.tech, this is one of the most frequently asked Spark/Big Data interview questions, reported at 1 company. DataEngPrep.tech maintains a curated database of 1,863+ real data engineering interview questions across 7 categories, verified by industry professionals.