Axon Server 2026.1.0: Multi-Tier DCB Storage and Self-Healing Topology

We're excited to announce the release of Axon Server 2026.1.0. This release is for anyone running Axon Server in production: platform and ops teams keeping clusters healthy through restarts and deployments, and developers relying on Axon Server’s event store – specifically its DCB (Dynamic Consistency Boundary) contexts – to hold years of event history without the fast-disk prices for all of it. The result: clusters have stable routing on their own, and event history stored at a lower cost, with nothing at risk of being lost.

Previously, topology could drift out of sync across the cluster. With this release, we’ve improved how topology information is distributed across the cluster, minimizing room for inconsistencies. And when routing disagreements do occur, the cluster reconciles its own routing state automatically.

Here's what that means for the clusters you run:

• Keep the history, cut the cost. DCB contexts can now age older segments onto cheaper storage automatically, so a long-lived event store isn't stuck entirely on your priciest disk.

• Upgrade live, node by node. The new communication layer is backward compatible, so you can move a cluster to 2026.1.0 node by node, with zero downtime.

• Nodes drift, the cluster corrects. Cluster routing now converges faster in cases of restarted nodes or infrastructure changes, checksum based topology sharing rules out accidental drift of routing entries.

What's New in Axon Server 2026.1.0

Multi-tier storage for DCB contexts

DCB contexts hold the full event history your application needs to stay correct, which made them expensive to run in very large deployments. Every byte sat on the same fast, expensive disk, regardless of age or how often anyone actually read it.

DCB (Dynamic Consistency Boundary) contexts now support tiered event storage:

• Writes land on the primary tier: fast local disk

• Sealed segments migrate automatically to cheaper or larger storage locations, driven by a size or age policy per tier

• An optional terminal black hole tier deletes what reaches it, to express retention

Keeping a large event store entirely on fast disk is expensive, and until now DCB contexts left no alternative. Tiers are configured per context and per node role. Configured per context - from the context settings in the Axon Server UI, or with the CLI.

You keep the full history a DCB context needs to stay correct, while only the data you actually touch sits on your most expensive disk. For a broader look at how Axon Server manages event storage, Axoniq Academy's Axon Server as an Event Store course is a good next step.

Rebuilt intra-cluster communication

If a subscription message got dropped during a restart or a rolling deployment, two nodes could end up disagreeing about where a message should go, with nothing forcing them back into sync.

Nodes now talk over a single bidirectional stream per pair of nodes, carrying all inter-node traffic, and they no longer push handler information subscription by subscription: each node periodically shares a checksum of its view on the topology, and a node that spots a mismatch pulls the peer's full topology and reconciles.

Now, the cluster converges on a correct shared view and repairs itself if it ever drifts, so routing stays accurate through restarts, network blips, and rolling deployments. It needs no configuration and is backward compatible: nodes that haven't been upgraded keep speaking the old protocol, so clusters upgrade node by node without downtime. 

For the full mechanics of cluster topology and node roles, Axoniq Academy's Managing an Axon Server Cluster course covers it in depth.

Upgrade to Axon Server 2026.1.0

Axon Server 2026.1.0 is available now on the download page, with full release notes covering every fix and improvement, including hardening around segment merging, snapshot retrieval, and backups. 

Have a question or feedback? Talk to us on the Discuss forum.

Go further with Axoniq Academy

If you want structured, hands-on training that goes deeper into Axon Server capabilities:

• Axon Server as an Event Store: how Axon Server stores, finds, and distributes events across a cluster

• Managing an Axon Server Cluster: cluster administration, node roles, and scaling in depth

• Deploying and Running Axon Server: planning deployments, backup strategy, and disaster recovery