Improve AT_LEAST_THREE least loaded node implementation.

What's Changed

• Allows more than one connecting nodes when AT_LEAST_THREE was used for least loaded node algorithm and connecting/connected nodes count is less than 3. by @Yellow-Rice in #533

Full Changelog: 2.4.1.81...2.4.1.82

3.0.1.80

What's Changed

• [LI-HOTFIX] Ensure ReplicationFactor is at least (minIsr + RedundancyFactor + 1) in topic creation by @hshi2022 in #532

Full Changelog: 3.0.1.79...3.0.1.80

3.0.1.882

What's Changed

• [LI-HOTFIX] Ensure ReplicationFactor is at least (minIsr + RedundancyFactor + 1) in topic creation by @hshi2022 in #532

Full Changelog: 3.0.1.79...3.0.1.882

2.4.1.81

What's Changed

• Replace expensive exception hashcode computation with ExceptionFingerprint by @sutambe in #529
• [LI-HOTFIX] throw RetriableKafkaClientConstructionException when resolved addresses is empty by @Q1Liu in #527

Full Changelog: 2.4.1.80...2.4.1.81

3.0.1.79

What's Changed

• Catch exceptions for LogDirFailureHandler thread during handleLogDirFailure by @CCisGG in #530

Full Changelog: 3.0.1.78...3.0.1.79

3.0.1.78

What's Changed

• Remove shutdown hook in KafkaYammerMetrics by @CCisGG in #528
• Fix logging order issue in GroupCoordinator by @CCisGG in #526

Full Changelog: 3.0.1.77...3.0.1.78

[LI-HOTFIX] Reduce exception log spam in producer IO thread

Fixes LIKAFKA-62226

Certain issues, such as DNS resolution failures, can cause the producer IO thread to spew log messages rapidly. This can cause disks to fill and hide other issues. This commit addresses this issue by adding a producer configuration that throttles exception log messages from the top of the producer IO thread. The throttling is a simple cooldown-based rate limiter, and is applied per-exception. Exceptions are differentiated based on their printStackTrace result, which covers exception type, construction point, causes, etc.

3.0.1.77

What's Changed

• checking newNode size to make sure its not 0 by @JobseRyan in #520
• Add LiMinSegmentRollMs to prevent log roll to be so fast by @CCisGG in #521

Full Changelog: 3.0.1.75...3.0.1.76

What's Changed

• LIKAFKA-62136: Change log level for client quota metric to debug by @allenxwang in #524

New Contributors

• @allenxwang made their first contribution in #524

Full Changelog: 2.4.1.76...3.0.1.77

2.4.1.79

What's Changed

• ACTIONITEM-3738: Improve the log message for the "There are no nodes in the Kafka cluster" error. by @Yellow-Rice in #523

Full Changelog: 2.4.1.77...2.4.1.79

2.4.1.78

checking new node size

What's Changed

• checking that newNodes size is not 0 before passing it into random number generator by @JobseRyan in #522

Full Changelog: 2.4.1.75...2.4.1.78