Kafka Exactly-Once Semantics Language Collocations
Practise the standard verbs for implementing Kafka exactly-once semantics.
0 / 5 completed
1 / 5
Fill in: 'We ___ idempotent producers so a retried send after a network blip never results in the same message being written twice.'
We 'enable' idempotence — the standard, established Kafka collocation for turning on deduplicated producer retries. The other options aren't the recognised term here.
2 / 5
Fill in: 'Using at-least-once delivery without idempotent consumers can ___ a duplicate message processed twice after a rebalance.'
We say at-least-once alone will 'leave' duplicates possible — the standard, natural collocation for the resulting risk. The other options aren't idiomatic here.
3 / 5
Fill in: 'We ___ transactions across the read-process-write cycle so a consumed offset and its produced output commit atomically together.'
We 'use' transactions — the standard, simple collocation for adopting Kafka's transactional API. The other options are less idiomatic here.
4 / 5
Fill in: 'We ___ a unique transactional ID per producer instance so fenced zombie producers can't write conflicting data after a restart.'
We 'assign an ID' — the standard, established Kafka collocation for giving a producer a stable transactional identifier. The other options aren't the recognised term here.
5 / 5
Fill in: 'We ___ duplicate processing rates before and after enabling exactly-once semantics to prove the change actually closed the gap.'
We 'compare' rates — the standard, simple collocation for contrasting metrics before and after a change. The other options are less idiomatic here.