Hot on the heels of Spring Framework 6.1, I am pleased to announce the 3.1 releases of Spring for Apache Kafka and Spring AMQP (Spring for RabbitMQ).

See the release notes for the 3.1.0, 3.1.0-RC1 and 3.1.0-M1 releases of each project - Spring for Apache Kafka and Spring AMQP for detailed content.

Also see the What's New documentation sections for highlights: Spring for Apache Kafka Spring AMQP.

I am pleased to announce that new major releases (3.0.0) for Spring for Apache Kafka and Spring for RabbitMQ are now available.

Common Changes

The 3.0.x versions of these projects have the following common changes

• they depend on Java version 17 or later

• they depend on Spring Framework 6.0.x

• support for creating native GraalVM applications

• support for micrometer observability and tracing

• Bill of Materials POMs are now provided to help with dependency management

Spring for Apache Kafka

• support for a global single EmbeddedKafkaBroker (across multiple test classes)

• several enhancements for @RetryableTopic (non-blocking retries)

• KafkaTemplate methods now return CompleteableFuture instead of ListenableFuture, which has been deprecated

…

I am pleased to announce that Spring for Apache Kafka 2.9.0 is now available.

This version is not provided by Spring Boot dependency management, but it can be used with Boot 2.7 (or 2.6), as long as you override the Kafka dependencies to 3.2.0 as described in this appendix, especially if you are using the embedded Kafka broker.

Please see the release candidate announcement for notable changes in this release.

For information about all changes in this release, see What’s New.

Project Page | GitHub | Issues | Documentation | Stack Overflow | Gitter

I am pleased to announce that the release candidate Spring for Apache Kafka 2.9.0 is now available (2.9.0-RC1) in the Spring Milestone Repo.

This version is not provided by Spring Boot dependency management, but it can be used with Boot 2.7 (or 2.6), as long as you override the Kafka dependencies as described in this appendix, especially if you are using the embedded Kafka broker.

Notable Changes

• This version uses the 3.2.0 kafka-clients version

• Non-blocking retry bootstrapping is now more robust

• New Error Handler Mode

By default, after an error, the DefaultErrorHandler performs seeks on the remaining records from the last poll and re-fetches them from the broker on the next poll. With high error rates and large max.poll.records, this can cause unnecessary strain on the network. For this reason, the error handler has a new property seekAfterError, when set to false, instead of seeking the records, the remaining records are retained in memory and the consumer paused for the next poll (or multiple polls if the error handler is configured to use the ContainerPausingBackOffHandler…

The recently released versions of Spring AMQP (2.4.0, 2.3.12, 2.2.20) contain a fix for CVE-2021-22095, which can cause a potential OutOfMemoryError for very large messages.

Spring Boot version 2.6.0 and 2.5.7 will bring in the 2.4.0 and 2.3.12 versions of spring-amqp respectively. Users of Boot 2.4.x should override the spring-amqp and spring-rabbit versions to 2.3.12; users of Spring Boot 2.3.x should override the versions to 2.2.20.

I am pleased to announce that Spring AMQP 2.4.0 is now available.

Notable Changes

• Initial support for the RabbitMQ Stream Plugin

• Support for remoting using Spring Framework’s RMI support is deprecated

For information about all changes in this release, see What’s New.

Project Page | GitHub | Issues | Documentation | Stack Overflow | Gitter

I am pleased to announce that Spring for Apache Kafka 2.8.0 is now available.

Notable Changes

• This version uses the 3.0.0 kafka-clients version

• Out of order manual commits are now supported

• The same container factory can now be used for both batch and record listeners

• Separate record and batch error handlers have been replaced with common error handlers

• KafkaTemplate can now be used to receive specific records from known partitions/offsets

For information about all changes in this release, see What’s New.

Project Page | GitHub | Issues | Documentation | Stack Overflow | Gitter

I am pleased to announce that Spring for Apache Kafka 2.7.0 is now available.

This release contains a significant enhancement, which is a community contribution. Failed deliveries can be forwarded to a series of topics for delayed redelivery.

It is best described with an example:

@RetryableTopic(attempts = "5", backoff = @Backoff(delay = 1000, multiplier = 2.0))
@KafkaListener(id = "sk270", topics = "sk270")
public void listen(String in, @Header(KafkaHeaders.RECEIVED_TOPIC) String topic) {
    LOG.info(in + " from " + topic);
    throw new RuntimeException("test");
}

@DltHandler
public void…

The 2.5.0.RC1 release candidate is now avaialable in the Spring milestone repo.

Update: 2.5.0.RELEASE was released on May 13th.

Highlights:

• kafka-clients 2.5.0 (alignment of version numbers is coincidental).

• Support for re-committing retryable offset commit exceptions for retained partitions when using cooperative rebalancing.

• Support for the new "fetch-offset-request" procuder fencing (when brokers are 2.5 or higher), requiring fewer producers.

• Support for static group membership.

• More integration with Micrometer.

• Optional Delivery Attempts header.

• RecoveringBatchErrorHandler can commit a partial batch and replay from failed record in a batch (with cooperation of the listener); this is now the default for a batch listener.

  …

To support last week’s Apache Kafka 2.4.0 release, I am pleased to announce that the Spring for Apache Kafka 2.4 - 2.4.0.RELEASE - is available in maven central.

This version is essentially functionally equivalent to 2.3.x, but is compiled against the 2.4.0 kafka-clients and supports the new incremental rebalancing protocol.

The 2.4.0 kafka-clients are not binary compatible with Spring for Apache Kafka 2.3 so if you wish to use the 2.4.0 clients, you must upgrade to this version. See the appendix in the reference manual for how to override the jar versions, especially if you are using Spring…