Page History

...

• Many internal optimisations in the Camel routing engine, such as reducing thread contention when updating JMX statistics, reducing internal state objects to claim less memory, and reducing the number of allocated objects to reduce overhead on GC etc, and much more. 
• The Camel Maven Plugin can now validate for duplicate route ids in your source code.
• Splitted Twitter component into 4, now directmessage, seach, streaming and timeline has its own endpoint and scheme. See documentation for more details
• Introduced HeadersMapFactory SPI which allows to plugin different implementations, or to use case sensitive maps that are faster than the default.
• Allow Kafka consumer to break on first unhandled exception, sync the offset from last known good, and then re-connect after one timeout cycle, to restart consuming again. This avoids loosing the failed message, but retry it again on either this consumer, or another consume which was re-balanced by Kafka. This requires to be turned on with the new option breakOnFirstError which can be set on both component or endpoint level.
• Starting and stoping the CamelContext when used with Spring framework (SpringCamelContext) was revised to ensure that the Camel context is started last - when all resources should be available, and stopped first - while all resources are still available
• Tracing via Tracer has been refactored to be using the Message History EIP which allows Camel to use same logic instead of doing twice the tracing/capture during runtime. This reduced the memory footprint.
• The SQL Stored Procedure now supports specifying custom types as FQN classnames and scale in numeric values.

...