Reactive Messaging RabbitMQ Connector Reference Documentation
This guide is the companion from the Getting Started with RabbitMQ. It explains in more details the configuration and usage of the RabbitMQ connector for reactive messaging.
| This documentation does not cover all the details of the connector. Refer to the SmallRye Reactive Messaging website for further details. |
The RabbitMQ connector allows Quarkus applications to send and receive messages using the AMQP 0.9.1 protocol. More details about the protocol can be found in the AMQP 0.9.1 specification.
| The RabbitMQ connector supports AMQP 0-9-1, which is very different from the AMQP 1.0 protocol used by the AMQP 1.0 connector. You can use the AMQP 1.0 connector with RabbitMQ as described in the AMQP 1.0 connector reference, albeit with reduced functionality. |
|
This technology is considered preview. In preview, backward compatibility and presence in the ecosystem is not guaranteed. Specific improvements might require changing configuration or APIs, and plans to become stable are under way. Feedback is welcome on our mailing list or as issues in our GitHub issue tracker. For a full list of possible statuses, check our FAQ entry. |
RabbitMQ connector extension
To use the connector, you need to add the quarkus-messaging-rabbitmq extension.
You can add the extension to your project using:
> ./mvnw quarkus:add-extensions -Dextensions="quarkus-messaging-rabbitmq"Or just add the following dependency to your project:
<dependency>
<groupId>io.quarkus</groupId>
<artifactId>quarkus-messaging-rabbitmq</artifactId>
</dependency>Once added to your project, you can map channels to RabbitMQ exchanges or queues by configuring the connector attribute:
# Inbound
mp.messaging.incoming.[channel-name].connector=smallrye-rabbitmq
# Outbound
mp.messaging.outgoing.[channel-name].connector=smallrye-rabbitmqoutgoing channels are mapped to RabbitMQ exchanges and incoming channels are mapped to RabbitMQ queues as required
by the broker.
Configuring the RabbitMQ Broker access
The RabbitMQ connector connects to RabbitMQ brokers.
To configure the location and credentials of the broker, add the following properties in the application.properties:
rabbitmq-host=amqp (1)
rabbitmq-port=5672 (2)
rabbitmq-username=my-username (3)
rabbitmq-password=my-password (4)
mp.messaging.incoming.prices.connector=smallrye-rabbitmq (5)| 1 | Configures the broker host name. You can do it per channel (using the host attribute) or globally using rabbitmq-host |
| 2 | Configures the broker port. You can do it per channel (using the port attribute) or globally using rabbitmq-port. The default is 5672. |
| 3 | Configures the broker username if required. You can do it per channel (using the username attribute) or globally using rabbitmq-username. |
| 4 | Configures the broker password if required. You can do it per channel (using the password attribute) or globally using rabbitmq-password. |
| 5 | Instructs the prices channel to be managed by the RabbitMQ connector |
In dev mode and when running tests, Dev Services for RabbitMQ automatically starts a RabbitMQ broker.
Receiving RabbitMQ messages
Let’s imagine your application receives Message<Double>.
You can consume the payload directly:
package inbound;
import org.eclipse.microprofile.reactive.messaging.Incoming;
import jakarta.enterprise.context.ApplicationScoped;
@ApplicationScoped
public class RabbitMQPriceConsumer {
@Incoming("prices")
public void consume(double price) {
// process your price.
}
}Or, you can retrieve the Message<Double>:
package inbound;
import org.eclipse.microprofile.reactive.messaging.Incoming;
import org.eclipse.microprofile.reactive.messaging.Message;
import jakarta.enterprise.context.ApplicationScoped;
import java.util.concurrent.CompletionStage;
@ApplicationScoped
public class RabbitMQPriceMessageConsumer {
@Incoming("prices")
public CompletionStage<Void> consume(Message<Double> price) {
// process your price.
// Acknowledge the incoming message, marking the RabbitMQ message as `accepted`.
return price.ack();
}
}Inbound Metadata
Messages coming from RabbitMQ contain an instance of IncomingRabbitMQMetadata in the metadata.
Optional<IncomingRabbitMQMetadata> metadata = incoming.getMetadata(IncomingRabbitMQMetadata.class);
metadata.ifPresent(meta -> {
final Optional<String> contentEncoding = meta.getContentEncoding();
final Optional<String> contentType = meta.getContentType();
final Optional<String> correlationId = meta.getCorrelationId();
final Optional<ZonedDateTime> creationTime = meta.getCreationTime(ZoneId.systemDefault());
final Optional<Integer> priority = meta.getPriority();
final Optional<String> replyTo = meta.getReplyTo();
final Optional<String> userId = meta.getUserId();
// Access a single String-valued header
final Optional<String> stringHeader = meta.getHeader("my-header", String.class);
// Access all headers
final Map<String,Object> headers = meta.getHeaders();
// ...
});Deserialization
The connector converts incoming RabbitMQ Messages into Reactive Messaging Message<T> instances. The payload type T depends on the value of the RabbitMQ received message Envelope content_type and content_encoding properties.
| content_encoding | content_type | T |
|---|---|---|
Value present |
n/a |
|
No value |
|
|
No value |
|
a JSON element which can be a |
No value |
Anything else |
|
If you send objects with this RabbitMQ connector (outbound connector), they are encoded as JSON and sent with content_type set to application/json. You can receive this payload using (Vert.x) JSON Objects, and then map it to the object class you want:
@ApplicationScoped
public static class Generator {
@Outgoing("to-rabbitmq")
public Multi<Price> prices() { (1)
AtomicInteger count = new AtomicInteger();
return Multi.createFrom().ticks().every(Duration.ofMillis(1000))
.map(l -> new Price().setPrice(count.incrementAndGet()))
.onOverflow().drop();
}
}
@ApplicationScoped
public static class Consumer {
List<Price> prices = new CopyOnWriteArrayList<>();
@Incoming("from-rabbitmq")
public void consume(JsonObject p) { (2)
Price price = p.mapTo(Price.class); (3)
prices.add(price);
}
public List<Price> list() {
return prices;
}
}| 1 | The Price instances are automatically encoded to JSON by the connector |
| 2 | You can receive it using a JsonObject |
| 3 | Then, you can reconstruct the instance using the mapTo method |
The mapTo method uses the Quarkus Jackson mapper. Check this guide to learn more about the mapper configuration.
|
Acknowledgement
When a Reactive Messaging Message associated with a RabbitMQ Message is acknowledged, it informs the broker that the message has been accepted.
Whether you need to explicitly acknowledge the message depends on the auto-acknowledgement setting for the channel; if that is set to true then your message will be automatically acknowledged on receipt.
Failure Management
If a message produced from a RabbitMQ message is nacked, a failure strategy is applied. The RabbitMQ connector supports three strategies, controlled by the failure-strategy channel setting:
-
fail- fail the application; no more RabbitMQ messages will be processed. The RabbitMQ message is marked as rejected. -
accept- this strategy marks the RabbitMQ message as accepted. The processing continues ignoring the failure. -
reject- this strategy marks the RabbitMQ message as rejected (default). The processing continues with the next message.
Sending RabbitMQ messages
Serialization
When sending a Message<T>, the connector converts the message into a RabbitMQ Message. The payload is converted to the RabbitMQ Message body.
| T | RabbitMQ Message Body |
|---|---|
primitive types or |
String value with |
Serialized String payload with |
|
|
Binary content, with |
|
Binary content, with |
Any other class |
The payload is converted to JSON (using a Json Mapper) then serialized with |
If the message payload cannot be serialized to JSON, the message is nacked.
Outbound Metadata
When sending Messages, you can add an instance of OutgoingRabbitMQMetadata
to influence how the message is handled by RabbitMQ. For example, you can configure the routing key, timestamp and
headers:
final OutgoingRabbitMQMetadata metadata = new OutgoingRabbitMQMetadata.Builder()
.withHeader("my-header", "xyzzy")
.withRoutingKey("urgent")
.withTimestamp(ZonedDateTime.now())
.build();
// Add `metadata` to the metadata of the outgoing message.
return Message.of("Hello", Metadata.of(metadata));Configuring the RabbitMQ Exchange/Queue
You can configure the RabbitMQ exchange or queue associated with a channel using properties on the channel configuration.
incoming channels are mapped to RabbitMQ queues and outgoing channels are mapped to RabbitMQ exchanges.
For example:
mp.messaging.incoming.prices.connector=smallrye-rabbitmq
mp.messaging.incoming.prices.queue.name=my-queue
mp.messaging.outgoing.orders.connector=smallrye-rabbitmq
mp.messaging.outgoing.orders.exchange.name=my-order-queueIf the exchange.name or queue.name attribute is not set, the connector uses the channel name.
To use an existing queue, you need to configure the name and set the exchange’s or queue’s declare property to false.
For example, if you have a RabbitMQ broker configured with a people exchange and queue, you need the following configuration:
mp.messaging.incoming.people.connector=smallrye-rabbitmq
mp.messaging.incoming.people.queue.name=people
mp.messaging.incoming.people.queue.declare=false
mp.messaging.outgoing.people.connector=smallrye-rabbitmq
mp.messaging.outgoing.people.exchange.name=people
mp.messaging.outgoing.people.exchange.declare=falseExecution model and Blocking processing
Reactive Messaging invokes your method on an I/O thread.
See the Quarkus Reactive Architecture documentation for further details on this topic.
But, you often need to combine Reactive Messaging with blocking processing such as database interactions.
For this, you need to use the @Blocking annotation indicating that the processing is blocking and should not be run on the caller thread.
For example, The following code illustrates how you can store incoming payloads to a database using Hibernate with Panache:
import io.smallrye.reactive.messaging.annotations.Blocking;
import org.eclipse.microprofile.reactive.messaging.Incoming;
import jakarta.enterprise.context.ApplicationScoped;
import jakarta.transaction.Transactional;
@ApplicationScoped
public class PriceStorage {
@Incoming("prices")
@Blocking
@Transactional
public void store(int priceInUsd) {
Price price = new Price();
price.value = priceInUsd;
price.persist();
}
}|
There are 2
They have the same effect. Thus, you can use both. The first one provides more fine-grained tuning such as the worker pool to use and whether it preserves the order. The second one, used also with other reactive features of Quarkus, uses the default worker pool and preserves the order. |
|
@RunOnVirtualThread
For running the blocking processing on Java virtual threads, see the Quarkus Virtual Thread support with Reactive Messaging documentation. |
Customizing the underlying RabbitMQ client
The connector uses the Vert.x RabbitMQ client underneath. More details about this client can be found in the Vert.x website.
You can customize the underlying client configuration by producing an instance of RabbitMQOptions as follows:
@Produces
@Identifier("my-named-options")
public RabbitMQOptions getNamedOptions() {
PemKeyCertOptions keycert = new PemKeyCertOptions()
.addCertPath("./tls/tls.crt")
.addKeyPath("./tls/tls.key");
PemTrustOptions trust = new PemTrustOptions().addCertPath("./tlc/ca.crt");
// You can use the produced options to configure the TLS connection
return new RabbitMQOptions()
.setSsl(true)
.setPemKeyCertOptions(keycert)
.setPemTrustOptions(trust)
.setUser("user1")
.setPassword("password1")
.setHost("localhost")
.setPort(5672)
.setVirtualHost("vhost1")
.setConnectionTimeout(6000) // in milliseconds
.setRequestedHeartbeat(60) // in seconds
.setHandshakeTimeout(6000) // in milliseconds
.setRequestedChannelMax(5)
.setNetworkRecoveryInterval(500) // in milliseconds
.setAutomaticRecoveryEnabled(true);
}This instance is retrieved and used to configure the client used by the connector.
You need to indicate the name of the client using the client-options-name attribute:
mp.messaging.incoming.prices.client-options-name=my-named-optionsTLS Configuration
RabbitMQ Messaging extension integrates with the Quarkus TLS registry to configure the Vert.x RabbitMQ client.
To configure the TLS for a channel, you need to provide a named TLS configuration in the application.properties:
quarkus.tls.your-tls-config.trust-store.pem.certs=ca.crt,ca2.pem
# ...
mp.messaging.incoming.prices.tls-configuration-name=your-tls-configHealth reporting
If you use the RabbitMQ connector with the quarkus-smallrye-health extension, it contributes to the readiness and liveness probes.
The RabbitMQ connector reports the readiness and liveness of each channel managed by the connector.
To disable health reporting, set the health-enabled attribute for the channel to false.
On the inbound side (receiving messages from RabbitMQ), the check verifies that the receiver is connected to the broker.
On the outbound side (sending records to RabbitMQ), the check verifies that the sender is not disconnected from the broker; the sender may still be in an initialised state (connection not yet attempted), but this is regarded as live/ready.
Note that a message processing failures nacks the message, which is then handled by the failure-strategy.
It’s the responsibility of the failure-strategy to report the failure and influence the outcome of the checks.
The fail failure strategy reports the failure, and so the check will report the fault.
Dynamic Credentials
Quarkus and the RabbitMQ connector support Vault’s RabbitMQ secrets engine for generating short-lived dynamic credentials. This allows Vault to create and retire RabbitMQ credentials on a regular basis.
First we need to enable Vault’s rabbitmq secret engine, configure it with RabbitMQ’s connection and authentication
information, and create a Vault role my-role (replace 10.0.0.3 by the actual host that is running the
RabbitMQ container):
vault secrets enable rabbitmq
vault write rabbitmq/config/connection \
connection_uri=http://10.0.0.3:15672 \
username=guest \
password=guest
vault write rabbitmq/roles/my-role \
vhosts='{"/":{"write": ".*", "read": ".*"}}'|
For this use case, user |
Then we need to give a read capability to the Quarkus application on path rabbitmq/creds/my-role.
cat <<EOF | vault policy write vault-rabbitmq-policy -
path "secret/data/myapps/vault-rabbitmq-test/*" {
capabilities = ["read"]
}
path "rabbitmq/creds/my-role" {
capabilities = [ "read" ]
}
EOFNow that Vault knows how to create users in RabbitMQ, we need to configure Quarkus to use a credentials-provider for RabbitMQ.
First we tell Quarkus to request dynamic credentials using a credentials-provider named rabbitmq.
quarkus.rabbitmq.credentials-provider = rabbitmq
Next we configure the rabbitmq credentials provider. The credentials-role option must be set to the name of the
role we created in Vault, in our case my-role. The credentials-mount option must be set to rabbitmq.
quarkus.vault.credentials-provider.rabbitmq.credentials-role=my-role
quarkus.vault.credentials-provider.rabbitmq.credentials-mount=rabbitmq
The credentials-mount is used directly as the mount of the secret engine in Vault. Here we are using
the default mount path of rabbitmq. If the RabbitMQ secret engine was mounted at a custom path, the
credentials-mount option must be set to that path instead.
|
RabbitMQ Connector Configuration Reference
Incoming channel configuration
| Attribute (alias) | Description | Mandatory | Default |
|---|---|---|---|
username (rabbitmq-username) |
The username used to authenticate to the broker Type: string |
false |
|
password (rabbitmq-password) |
The password used to authenticate to the broker Type: string |
false |
|
host (rabbitmq-host) |
The broker hostname Type: string |
false |
|
port (rabbitmq-port) |
The broker port Type: int |
false |
|
ssl (rabbitmq-ssl) |
Whether the connection should use SSL Type: boolean |
false |
|
trust-all (rabbitmq-trust-all) |
Whether to skip trust certificate verification Type: boolean |
false |
|
trust-store-path (rabbitmq-trust-store-path) |
The path to a JKS trust store Type: string |
false |
|
trust-store-password (rabbitmq-trust-store-password) |
The password of the JKS trust store Type: string |
false |
|
credentials-provider-name (rabbitmq-credentials-provider-name) |
The name of the RabbitMQ Credentials Provider bean used to provide dynamic credentials to the RabbitMQ client Type: string |
false |
|
connection-timeout |
The TCP connection timeout (ms); 0 is interpreted as no timeout Type: int |
false |
|
handshake-timeout |
The AMQP 0-9-1 protocol handshake timeout (ms) Type: int |
false |
|
automatic-recovery-enabled |
Whether automatic connection recovery is enabled Type: boolean |
false |
|
automatic-recovery-on-initial-connection |
Whether automatic recovery on initial connections is enabled Type: boolean |
false |
|
reconnect-attempts (rabbitmq-reconnect-attempts) |
The number of reconnection attempts Type: int |
false |
|
reconnect-interval (rabbitmq-reconnect-interval) |
The interval (in seconds) between two reconnection attempts Type: int |
false |
|
network-recovery-interval |
How long (ms) will automatic recovery wait before attempting to reconnect Type: int |
false |
|
user |
The AMQP username to use when connecting to the broker Type: string |
false |
|
include-properties |
Whether to include properties when a broker message is passed on the event bus Type: boolean |
false |
|
requested-channel-max |
The initially requested maximum channel number Type: int |
false |
|
requested-heartbeat |
The initially requested heartbeat interval (seconds), zero for none Type: int |
false |
|
use-nio |
Whether usage of NIO Sockets is enabled Type: boolean |
false |
|
virtual-host (rabbitmq-virtual-host) |
The virtual host to use when connecting to the broker Type: string |
false |
|
exchange.name |
The exchange that messages are published to or consumed from. If not set, the channel name is used. If set to Type: string |
false |
|
exchange.durable |
Whether the exchange is durable Type: boolean |
false |
|
exchange.auto-delete |
Whether the exchange should be deleted after use Type: boolean |
false |
|
exchange.type |
The exchange type: direct, fanout, headers or topic (default) Type: string |
false |
|
exchange.declare |
Whether to declare the exchange; set to false if the exchange is expected to be set up independently Type: boolean |
false |
|
tracing.enabled |
Whether tracing is enabled (default) or disabled Type: boolean |
false |
|
tracing.attribute-headers |
A comma-separated list of headers that should be recorded as span attributes. Relevant only if tracing.enabled=true Type: string |
false |
`` |
queue.name |
The queue from which messages are consumed. Type: string |
true |
|
queue.durable |
Whether the queue is durable Type: boolean |
false |
|
queue.exclusive |
Whether the queue is for exclusive use Type: boolean |
false |
|
queue.auto-delete |
Whether the queue should be deleted after use Type: boolean |
false |
|
queue.declare |
Whether to declare the queue and binding; set to false if these are expected to be set up independently Type: boolean |
false |
|
queue.ttl |
If specified, the time (ms) for which a message can remain in the queue undelivered before it is dead Type: long |
false |
|
queue.single-active-consumer |
If set to true, only one consumer can actively consume messages Type: boolean |
false |
|
queue.x-queue-type |
If automatically declare queue, we can choose different types of queue [quorum, classic, stream] Type: string |
false |
|
queue.x-queue-mode |
If automatically declare queue, we can choose different modes of queue [lazy, default] Type: string |
false |
|
max-incoming-internal-queue-size |
The maximum size of the incoming internal queue Type: int |
false |
|
connection-count |
The number of RabbitMQ connections to create for consuming from this queue. This might be necessary to consume from a sharded queue with a single client. Type: int |
false |
|
auto-bind-dlq |
Whether to automatically declare the DLQ and bind it to the binder DLX Type: boolean |
false |
|
dead-letter-queue-name |
The name of the DLQ; if not supplied will default to the queue name with '.dlq' appended Type: string |
false |
|
dead-letter-exchange |
A DLX to assign to the queue. Relevant only if auto-bind-dlq is true Type: string |
false |
|
dead-letter-exchange-type |
The type of the DLX to assign to the queue. Relevant only if auto-bind-dlq is true Type: string |
false |
|
dead-letter-routing-key |
A dead letter routing key to assign to the queue; if not supplied will default to the queue name Type: string |
false |
|
dlx.declare |
Whether to declare the dead letter exchange binding. Relevant only if auto-bind-dlq is true; set to false if these are expected to be set up independently Type: boolean |
false |
|
dead-letter-queue-type |
If automatically declare DLQ, we can choose different types of DLQ [quorum, classic, stream] Type: string |
false |
|
dead-letter-queue-mode |
If automatically declare DLQ, we can choose different modes of DLQ [lazy, default] Type: string |
false |
|
failure-strategy |
The failure strategy to apply when a RabbitMQ message is nacked. Accepted values are Type: string |
false |
|
broadcast |
Whether the received RabbitMQ messages must be dispatched to multiple subscribers Type: boolean |
false |
|
auto-acknowledgement |
Whether the received RabbitMQ messages must be acknowledged when received; if true then delivery constitutes acknowledgement Type: boolean |
false |
|
keep-most-recent |
Whether to discard old messages instead of recent ones Type: boolean |
false |
|
routing-keys |
A comma-separated list of routing keys to bind the queue to the exchange Type: string |
false |
|
content-type-override |
Override the content_type attribute of the incoming message, should be a valid MIME type Type: string |
false |
|
max-outstanding-messages |
The maximum number of outstanding/unacknowledged messages being processed by the connector at a time; must be a positive number Type: int |
false |
Outgoing channel configuration
| Attribute (alias) | Description | Mandatory | Default |
|---|---|---|---|
automatic-recovery-enabled |
Whether automatic connection recovery is enabled Type: boolean |
false |
|
automatic-recovery-on-initial-connection |
Whether automatic recovery on initial connections is enabled Type: boolean |
false |
|
connection-timeout |
The TCP connection timeout (ms); 0 is interpreted as no timeout Type: int |
false |
|
default-routing-key |
The default routing key to use when sending messages to the exchange Type: string |
false |
`` |
default-ttl |
If specified, the time (ms) sent messages can remain in queues undelivered before they are dead Type: long |
false |
|
exchange.auto-delete |
Whether the exchange should be deleted after use Type: boolean |
false |
|
exchange.declare |
Whether to declare the exchange; set to false if the exchange is expected to be set up independently Type: boolean |
false |
|
exchange.durable |
Whether the exchange is durable Type: boolean |
false |
|
exchange.name |
The exchange that messages are published to or consumed from. If not set, the channel name is used. If set to Type: string |
false |
|
exchange.type |
The exchange type: direct, fanout, headers or topic (default) Type: string |
false |
|
handshake-timeout |
The AMQP 0-9-1 protocol handshake timeout (ms) Type: int |
false |
|
host (rabbitmq-host) |
The broker hostname Type: string |
false |
|
include-properties |
Whether to include properties when a broker message is passed on the event bus Type: boolean |
false |
|
max-inflight-messages |
The maximum number of messages to be written to RabbitMQ concurrently; must be a positive number Type: long |
false |
|
max-outgoing-internal-queue-size |
The maximum size of the outgoing internal queue Type: int |
false |
|
network-recovery-interval |
How long (ms) will automatic recovery wait before attempting to reconnect Type: int |
false |
|
password (rabbitmq-password) |
The password used to authenticate to the broker Type: string |
false |
|
port (rabbitmq-port) |
The broker port Type: int |
false |
|
reconnect-attempts (rabbitmq-reconnect-attempts) |
The number of reconnection attempts Type: int |
false |
|
reconnect-interval (rabbitmq-reconnect-interval) |
The interval (in seconds) between two reconnection attempts Type: int |
false |
|
requested-channel-max |
The initially requested maximum channel number Type: int |
false |
|
requested-heartbeat |
The initially requested heartbeat interval (seconds), zero for none Type: int |
false |
|
ssl (rabbitmq-ssl) |
Whether the connection should use SSL Type: boolean |
false |
|
tracing.attribute-headers |
A comma-separated list of headers that should be recorded as span attributes. Relevant only if tracing.enabled=true Type: string |
false |
`` |
tracing.enabled |
Whether tracing is enabled (default) or disabled Type: boolean |
false |
|
trust-all (rabbitmq-trust-all) |
Whether to skip trust certificate verification Type: boolean |
false |
|
trust-store-password (rabbitmq-trust-store-password) |
The password of the JKS trust store Type: string |
false |
|
trust-store-path (rabbitmq-trust-store-path) |
The path to a JKS trust store Type: string |
false |
|
credentials-provider-name (rabbitmq-credentials-provider-name) |
The name of the RabbitMQ Credentials Provider bean used to provide dynamic credentials to the RabbitMQ client Type: string |
false |
|
use-nio |
Whether usage of NIO Sockets is enabled Type: boolean |
false |
|
user |
The AMQP username to use when connecting to the broker Type: string |
false |
|
username (rabbitmq-username) |
The username used to authenticate to the broker Type: string |
false |
|
virtual-host (rabbitmq-virtual-host) |
The virtual host to use when connecting to the broker Type: string |
false |
|
