Product announcements

Axon Framework 4.13.5 (LTS) is now available. This release focuses on improving interoperability between Axon Framework 4 and Axon Framework 5 for teams migrating gradually.

Starting with Axon Framework 5.3.2, the connector sends an empty payload for void results instead of omitting it entirely. On the Axon Framework 4 side, this triggered SerializationException errors when handling void command results and void subscription query updates. Empty payloads are now treated the same as absent ones, letting void handlers work correctly across versions.

Additionally, this release includes dependency upgrades to keep the LTS branch current.

Full release notes are available here.

Axon Framework 4.13.4 (LTS) is now available. There are two important fixes in this release, further improving how Event Processors recover from specific error conditions and how live events are handled after a selective replay.

The first fix resolves a critical issue where an Event Processor could become stuck when specific errors occurred during commit or abort, blocking event processing entirely for affected applications. Now, these errors are caught explicitly, rolling them back cleanly and surfacing the failure so processing can recover.

The second fix ensures that handlers using @ReplayContext receive live events correctly once a selective replay concludes, matching the documented contract and keeping applications in a consistent state.

Axon Framework 4.13.4 (LTS) release also improves interoperability between Axon Framework 4 and Axon Framework 5 for teams migrating gradually. Plain String results now deserialize correctly across versions, and queries from Axon Framework 5 correctly find their handler on Axon Framework 4 even without a response type attached.

Additionally, this release also includes the following improvements:

• Subscription queries convert their initial result consistently, whether it arrives as a single item or a list.

• Tracing handles a null metadata value correctly when generating span attributes.

• The BOM module has been moved into this repository, keeping its version aligned with the rest of the LTS modules.
  
  

Full release notes are available here. 

Axon Server 2026.1.0 is now available, introducing multi-tier storage for DCB contexts and self-healing cluster topology.

Multi-tier storage enables DCB (Dynamic Consistency Boundary) contexts to migrate older event segments to lower-cost storage, allowing organizations to retain long term event history while reserving premium, high-performance disk for the actively used data. Cluster topology is now shared and reconciled automatically across nodes, ensuring routing remains accurate through restarts, network disruptions, and rolling deployments with no manual intervention required.

This release also includes:

• Zero-downtime upgrades: The rebuilt communication layer is backward compatible, enabling clusters to be upgraded node by node while remaining fully operational throughout.

• Reliability hardening: Improvements to segment merging, snapshot retrieval, and backup handling strengthen overall cluster and data reliability.

Release notes are available here, and the full announcement can be read here. Install Axon Server 2026.1.0 now!

Axon and Axoniq Framework 5.3 is now available, introducing multitenancy and distributed tracing for enterprise customers.

Multitenancy enables a single deployment to serve multiple tenants while keeping event streams, commands, and queries fully isolated. Onboarding a new tenant requires only a configuration change, allowing organizations to scale without expanding their infrastructure footprint. Distributed tracing provides visibility into a request as it moves across services and handlers, reducing the time required to diagnose production issues.

This release also includes:

• Self-checkpointing handlers: Handlers performing asynchronous work, such as writing to a search index, can report progress only once the write is confirmed durable, and a built-in reconciliation engine handles this logic automatically.

• Event transformation: A single stored event can be split into multiple events going forward, allowing teams to correct earlier modeling decisions without modifying historical data.

Release notes are available here, and the full announcement can be read here.