fvaleri commented on code in PR #20221:
URL: https://github.com/apache/kafka/pull/20221#discussion_r2230353006
##########
tools/src/main/java/org/apache/kafka/tools/ConsumerPerformance.java:
##########
@@ -134,8 +136,13 @@ private static void consume(KafkaConsumer<byte[], byte[]>
consumer,
long reportingIntervalMs = options.reportingIntervalMs();
boolean showDetailedStats = options.showDetailedStats();
SimpleDateFormat dateFormat = options.dateFormat();
- consumer.subscribe(options.topic(),
- new ConsumerPerfRebListener(joinTimeMs, joinStartMs,
joinTimeMsInSingleRound));
+
+ ConsumerPerfRebListener listener = new
ConsumerPerfRebListener(joinTimeMs, joinStartMs, joinTimeMsInSingleRound);
Review Comment:
I see what you mean, but isn't that the same dynamic partition assignment
mechanism that we have when subscribing using a list of topics?
My understanding is that `fetchTimeInMs` can only be trusted when subscribed
partitions remains stable during the entire test run, or when you are using
classic rebalance protocol with a non cooperative assignment strategy. This
makes sense, because this tool was created when stop-the-world rebalancing was
the only option.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
