At-least-once delivery in Kafka - Cheat Sheet & Quick Revision

Recall & Review

beginner

What does At-least-once delivery mean in Kafka?

It means each message is delivered to the consumer one or more times, ensuring no message is lost but duplicates can occur.

intermediate

How does Kafka achieve At-least-once delivery?

By committing offsets only after processing messages, so if a failure happens before commit, messages are reprocessed.

beginner

What is the risk of At-least-once delivery?

The risk is duplicate message processing, because messages may be delivered more than once.

intermediate

Which Kafka consumer setting helps control At-least-once delivery behavior?

The enable.auto.commit setting controls if offsets are committed automatically or manually, affecting delivery guarantees.

beginner

Why might you choose At-least-once delivery over At-most-once?

Because it guarantees no message loss, which is important for critical data even if duplicates happen.

What does At-least-once delivery guarantee in Kafka?

A Messages may be lost but never duplicated

B No message is lost, but duplicates can occur

C Each message is delivered exactly once

D Messages are delivered only once without fail

At-least-once delivery ensures messages are never lost but may be delivered multiple times.

Which Kafka feature helps avoid message loss in At-least-once delivery?

A Auto-deleting messages immediately

B Disabling consumer groups

C Manual offset commit after processing

D Using only one partition

Manual offset commit after processing ensures messages are reprocessed if failure occurs before commit.

What is a common side effect of At-least-once delivery?

A Duplicate message processing

B Message loss

C Faster message delivery

D No need to commit offsets

Duplicates happen because messages may be delivered again if offsets are not committed.

If enable.auto.commit is set to true, what happens?

A Consumer crashes immediately

B Offsets are never committed

C Messages are always duplicated

D Offsets are committed automatically, possibly causing At-most-once delivery

Auto commit can cause offsets to be committed before processing, risking message loss (At-most-once).

Why is At-least-once delivery preferred for critical data?

A It prevents message loss even if duplicates occur

B It guarantees no duplicates

C It is faster than other methods

D It requires no offset management

Critical data needs no loss, so At-least-once is safer despite duplicates.

Explain how Kafka ensures At-least-once delivery and what trade-offs it involves.

Think about when offsets are saved and what happens if a failure occurs.

Describe the difference between At-least-once and At-most-once delivery in Kafka.

Consider when offsets are committed relative to message processing.