Enabling deduplication can impact broker performance, especially during informational snapshots, as it requires the broker to track and periodically snapshot sequence IDs for each producer. It also increases memory usage, since the broker must store deduplication information for active and recently inactive producers, and may lengthen topic recovery time after broker restarts or failovers. See
deduplication cookbook.
Would you like more detail on specific performance or operational impacts?
Sources:
-
Pulsar deduplication cookbook 👍 0 button 👎 0 button 💡
Hint: Mention @runllm in the thread for followups.