...
- The route models (definition classes) preserve property placeholders, which allows round trip editing, and also prevents showing potential sensitive information
- Added onPrepareFailure option to Dead Letter Channel to allow a custom processor to prepare the Exchange before sending to the dead letter queue.
- MongoDB tailable consumers now allow setting the
readPreference
option. - Using setFaultBody / setFaultHeader behaves similar to setBody / setHeader to preserve existing headers/attachments by setting on existing IN or OUT message.
- Java DSL using choice with end vs endChoice now validates you use them as intended. Any misconfiguration now throws an exception explaining the problem.
- POJO Producing and POJO Consuming lifecycle is now in sync with CamelContext which ensures it works as similar to Camel Routes does. Also the created producer/consumer is stopped when CamelContext is stopping to cleanup properly.
- JMS component now allows configuring concurrentConsumers/maxConcurrentConsumers separately between from(jms) vs to(jms) using request/reply over JMS
- Added SEARCH operation to Elasticsearch component
- The Properties component (property placeholder) now supports an encoding option to allow reading the .properties files using a specific charset such as UTF-8. By default ISO-8859-1 is used (latin1)
- Added
receviveBufferSize
option to the FTP component to make download faster. The size is by default 32kb buffer. - The Aggregator2 EIP now allows an external source using a
AggregateController
to control the aggregator, such as for forcing completing a group or all groups, or query runtime information from the aggregator. A default controller is provided out of the box if none has been configured. - The Aggregator2 now enlists in JMX using a
ManagedAggregateProcessorMBean
which has more information, and also allows using the aggregate controller to control it. - The Aggregator2 now includes more runtime statistics which can be accessed from Java and JMX about number of incoming and completed exchanges, and per different completion triggers.
- The Aggregator2 now supports pre-completion mode if the aggregation strategy implements
PreCompletionAwareAggregationStrategy
. This supports the use-case where an incoming Exchange are used to determine if the current aggregated group is completed or not. If completed the group is emitted as-is, and the new incoming Exchange is used to start a new group from scratch. - Added options to Mail consumer so it can skip or handle a mail message that cannot be read from the mail server in the batch; and then move on to read the next message. This ensures the consumer can deal with mail boxes that has problematic mails that otherwise will cause no mails to be retrieve by the consumer in the batch.
- The runtime route processors and their corresponding model definition are now linked by their id using
HasId
, so its easier to obtain the processor or model definition from thegetProcessor(id)
andgetProcessorDefintion(id)
api on CamelContext. - CamelContext now also includes a getManagedProcessor API to easily get the mbean client api for the processor, so you can access the runtime JMX statistics from Java code.
- Spring Boot applications now block the main thread of the execution
- Spring Boot automagically loads XML routes definitions from classpath
- Added
MainListener
to make it easier for Camel Boot / Standalone Camel to implement custom logic to trigger during the starting and stopping of the application. - Camel Jackson now provides
Map
toObject
converter - Ability to refer to existing hazelcast instance by id, using Hazelcast Component
- Direct and Direct-VM now supports the failIfNoConsumers option
- Making it easier to add custom Type Converter to your Camel applications without having to use the META-INF file. In Spring or Blueprint just add them as <bean>s. See more details at Type Converter documentation.
- camel-cxf PAYLOAD supports stream cache out of box.
BridgePropertyPlaceholderConfigurer
supports to handle the jasypt encryption inside and outside of the camelContext.The camel-aws component now support Amazon DynamoDB version 2
The Wire Tap eip now also emits event notifications when sending the Exchange to the wire tap destination
- Camel Test has new option isDumpRouteStats that dumps all route usage statistics as XML files in target/camel-route-stats directory, which can be used to inspect "route coverage" reports.
- Now Camel Netty component doesn't intercept consumers with
httpMethodRestrict=OPTIONS
- Now Netty HTTP resolves OPTIONS prefix matches earlier
- Using Load Balancer with Java DSL now builds the route model using the specific model types, which renders a more accurate route model
- Enhanced the PredicateBuilder to support vargs and list in the "or" builder method
- Using custom
HttpBinding
with HTTP and HTTP4 is now easier as they can be instantiated using a default no-arg constructor - Added
readLock=idempotent
to the File2 consumer which allows to reuse an idempotent repository as read-lock which enables Camel to run reliable in clustered mode, where each node competes for the same files from a shared file system. This requires that the idempotent repository supports clustering such as Hazelcast Component or Infinispan - or using the JDBC for a shared database table. - Camel RX now ensures that observe and subscribe are done within an unit of work to ensure on completions is executed, such as they are done when you use regular Camel routes.
- The JMS component now includes the underlying
javax.jms.Session
instance on the Camel message which allows end users to easier to use client acknowledge mode, in case the client wants to use the session to call recover or rollback. - Added
completionEager
option to Idempotent Consumer to whether to complete the idempotent consumer eager or when the exchange is done. - Working with the Camel source code in Eclipse with m2eclipse plugin improved to avoid endless file changed detection and recompile the source.
- Add option
suppressExceptions
to JSonPath which allows to ignore exceptions such as NoSuchPathException when the json payload does not contain the path defined in the json expression. - Spring Boot now creates
CamelContext
,ProducerTemplate
andConsumerTemplate
beans only if those are not defined by the user. - New
camel-archetype-spring-boot
to create a new Camel Spring Boot project using Maven. - New
camel-archetype-cdi
to create a new Camel CDI project using Maven. - FTP2 doesn't support absolute directories in endpoint URIs. In such cases, paths will be treated as relative and WARN message will be printed in logs. In future versions, there'll be no such assumption and validation exception will be thrown instead.
- Swagger module now supports getting api-docs from multiple camel contexts in the JVM. The entry point at /api-docs now lists the contexts detected, and you need to append the context id in the path, eg /api-docs/myCamel
- FTP2 supports
autoCreate=true
for ftp consumers - Rest DSL now supports to define parameter using fine grained details for the Swagger api documentation. For an example see the
camel-example-servlet-rest-tomcat
in the Apache Camel distribution. - The Simple language now has a equals ignore case operator (=~) when comparing with string values
- If using <contextScan> with Spring or Blueprint to filter RouteBuilder classes, then Camel will now by default only look for singleton beans. You can turn on the old behavior to include prototype scoped with the new option
includeNonSingletons
- The IdempotentRepository now supports clear operation, to clear the repository, and expose it over JMX.
- Netty4 in clientMode will now automatic re-connect if the connection was disconnected to the server.
- The
throwException
EIP now supports creating a new exception instance from a simple language expressions, which allows to make it dynamic based on information from the current Exchange. - When using Expression or Predicate in the routes, you can refer to the value as an external resource by using resource:classpath:path or resource:file:path - for example
resource:classpath:com/foo/myscript.groovy
. - Added option useMessageBodyForSql to the SQL Component which allows to use more dynamic SQL queries instead of the fixed SQL defined in the endpoint uri.
- Added sticky option to failover Load Balancer so the load balancer can start from the last known good endpoint.
- Added option
shareUnitOfWork
to Content Enricher eip - Content Enricher with enrich and pollEnrich now supports dynamic endpoint uris computed using an Expression that allows to use values from the current Exchange so you for example can poll a file which name is computed from data from the Exchange etc. This change breaks the XML DSL a little bit and there is an easy migration effort. The Java DSL stays backwards compatible.
- The Camel Karaf commands about routes now no longer require the context-id as 2nd parameter that is optional - if none provided then the command is a bulk operation for all Camel contexts.
- A simpler version of dynamic to with the new <toD> and toD in the DSL, documented at Message Endpoint
- Wire Tap now supports dynamic endpoint uris, in similar way the new <toD> and toD does
- Rest DSL now supports the new <toD> and toD in the rest-dsl.
- The EIP's now have more specialized processor mbeans in Camel JMX and you can also dump the processor as xml source to represent how it would appear in xml dsl
- Configuring Camel JMX using JVM system properties now all take precedence over any existing configuration (some option was not doing that)
- Introduced Extended JMX statistics level to gather additional metrics during runtime.
Fixed the following issues:
...
org.apache.camel.mode.LoadBalancerDefinition no longer implements org.apache.camel.processor.loadbalancer.LoadBalancer which is the runtime processor (this was never intended).
- The
ref
attribute on<loadBalance>
has been removed, as it has been deprecated for a long time, and you should use a<customLoadBalancer>
to refer to a custom load balancer. - The
copy
method onStreamCache
now takes anExchange
as parameter. - Various APIs in
camel-jms
has been adjusted to support including the JMS session parameter -javax.jms.Session
. Though these API changes are mostly internal facing, and we do not expect end users problems. - The
resourceUri
andresourceRef
attributes on<enrich>
and<pollEnrich>
has been removed as they now support a dynamic uris computed from an Expression. - Various APIs from camel-http in the package
org.apache.camel.component.http
has been moved to the camel-http-common module in the packageorg.apache.camel.http.common
which mean you may need to change the imports. - Renamed
All
enum onorg.apache.camel.ManagementStatisticsLevel
toDefault
Known Issues
Dependency upgrades
...
- The dumpRoutesAsXml operation now preserve property placeholder used in the route models.
- Using setFaultBody / setFaultHeader behaves similar to setBody / setHeader to preserve existing headers/attachments by setting on existing IN or OUT message.
- If using concurrent consumer on JMS endpoints for request/reply over JMS then you must use the new replyToConcurrentConsumers, replyToMaxConcurrentConsumers options to configure the values.
- When the Aggregator2 is forced completed the exchange property
Exchange.AGGREGATED_COMPLETED_BY
value has been changed from "forceCompletion" to "force" so its named like the other completion triggers. - Removed unsupported modules camel-web and camel-web-standalone.
- Removed unsupported camel:dot functionality from camel:run plugin.
- Removed unsupported camel-archetype-scala-component from maven archetypes.
- The Maven coordinate for linkedin and olingo2 components changed to be aligned like all the others having groupId as just org.apache.camel.
- If using MongoDB component, the option invokeGetLastError doesn't exist anymore. The option invokeGetLastError instructs the MongoDB Java driver to invoke getLastError() after every call. In the 3.0.0 MongoDB driver implementation, the getLastError() method in WriteResult has been removed. So, if needed, use an acknowledged WriteConcern when executing the write operation and then verify the correctness of the operation with the method wasAcknowledged() of WriteResult.
- The Jing component now uses
jing
as schema name in uris, instead ofrng
orrnc
. Those names has been removed. - Swagger module now supports getting api-docs from multiple camel contexts in the JVM. The entry point at /api-docs now lists the contexts detected, and you need to append the context id in the path, eg /api-docs/myCamel
- If using <contextScan> with Spring or Blueprint to filter RouteBuilder classes, then Camel will now by default only look for singleton beans. You can turn on the old behavior to include prototype scoped with the new option
includeNonSingletons
- camel-vertx has been upgraded to vertx 3.0 which requires Java 8 at runtime.
camel-cdi
is now using CDI 1.1 api - support for 1.0 has been dropped.- Content Enricher with enrich and pollEnrich now supports dynamic endpoint uris computed using an Expression that allows to use values from the current Exchange so you for example can poll a file which name is computed from data from the Exchange etc. This change breaks the XML DSL a little bit and there is an easy migration effort. The Java DSL stays backwards compatible.
- Support for CXF 2.7 is dropped.
- camel-jetty8 is being considered deprecated and we would like to drop this in a nearby future release as Jetty 8.x is EOL.
- If you have explicit configured the JMX statistics level to All then that option is now called Default.
Getting the Distributions
...