Fixes: #3090
Issue: https://github.com/spring-cloud/spring-cloud-stream/issues/3090
Fixes issue where KafkaBinderConfiguration would add customizers after the
transaction manager was already created.
The KafkaMessageChannelBinder was initializing the transaction manager in the
constructor before client factory customizers were added, which meant the customizers
were never applied to the producer factory used by the transaction manager.
- Moves transaction manager initialization from the constructor to the onInit() method
- Leverages the existing InitializingBean lifecycle to ensure customizers are applied
before the transaction manager is created
- Adds comprehensive test coverage to verify the fix in the binder and via auto-configuration
Signed-off-by: Soby Chacko <soby.chacko@broadcom.com>
Resolves#3091
- Producer that fetches partition info now initializes within a try-with-resources block
- If exceptions occur on calling producer.partitionsFor(topicName), it's now properly closed and resources are released
Signed-off-by: akenra <37288280+akenra@users.noreply.github.com>
Fixes: https://github.com/spring-cloud/spring-cloud-stream/issues/3062
The `KafkaBinderMetrics` creates `KafkaConsumer` instances and schedule the fix rate
task for them, but never closes them even when the `scheduler` is shut downed
* Implement a `Lifecycle` contract in the `KafkaBinderMetrics` and call `close()`
from the `stop()` to satisfy CRaC resource management expectations.
* Also close all the `KafkaConsumer` instances from the `metadataConsumers`
**Cherry-pick to `4.1.x`**
Fixes: #3040Resolves#3041
This change adds out-of-the-box `ListenerContainerCustomizer<AbstractMessageListenerContainer>`
to set `VirtualThreadTaskExecutor` into an `AbstractMessageListenerContainer` created by the binder
when `Threading.VIRTUAL` condition is met
Fixes: https://github.com/spring-cloud/spring-cloud-stream/issues/2939
The RabbitMQ 4.0 does not deal with client side `x-*` headers.
Therefore, an `x-death.count` is not incremented anymore when
message is re-published from client back to the broker.
* Spring AMQP 3.2 has introduced an `AmqpHeaders.RETRY_COUNT` custom header.
Use `messageProperties.incrementRetryCount()` in the `RabbitMessageChannelBinder`
when we re-published message back to the broker for server-side retries
* Fix docs respectively
Resolves#3019
Resolves https://github.com/spring-cloud/spring-cloud-stream/issues/2985
This commit introduces new customization options for Kafka listener containers
in Spring Cloud Stream, along with comprehensive documentation:
- Add KafkaListenerContainerCustomizer interface for Kafka-specific customization
with access to extended consumer properties
- Extend ListenerContainerWithDlqAndRetryCustomizer to include access to
extended consumer properties
- Update KafkaMessageChannelBinder to support the new customizer interfaces
- Implement KafkaListenerContainerCustomizerTests for integration testing
- Add detailed AsciiDoc reference documentation explaining the purpose,
usage, and hierarchy of these customizer interfaces:
* ListenerContainerCustomizer (existing)
* KafkaListenerContainerCustomizer (new)
* ListenerContainerWithDlqAndRetryCustomizer (extended)
- Update navigation to include the new documentation
These changes enhance the flexibility and configurability of Kafka consumer
endpoints in Spring Cloud Stream applications, allowing users to fine-tune
their listener containers based on specific requirements and scenarios,
with improved access to Kafka-specific properties.
- Integrate KafkaConnectionDetails, a Spring Boot component, in binder
- Update KafkaBinderConfigurationProperties to use KafkaConnectionDetails
- Modify KafkaTopicProvisioner to leverage KafkaConnectionDetails
- Adjust Kafka binder configurations to pass KafkaConnectionDetails
- Update tests to accommodate KafkaConnectionDetails changes
- Add KafkaConnectionDetails to shared.beans for auto-configuration
This change improves flexibility in configuring Kafka connections,
allowing for better support of externalized configuration management
and aligning with Spring Boot's connection abstraction model.
Fixes https://github.com/spring-cloud/spring-cloud-stream/issues/2650
* Enable native observability support for output binding in the reactive Kafka binder
* Adding test to verify this support with downstream consumers
* Adding ref docs
* Addressing PR review
Spring for Apache Pulsar introduced support for default tenant and namespace
for Pulsar topics in 6d23378fbb.
This ensures that all topic names are fully-qualified (using the default
tenant and namespace when not fully-qualified).
