DEVOPSdigest

Kurrent announced the release of KurrentDB 25.1, delivering its most developer-friendly event sourcing database yet and removing traditional barriers that have made event-driven development complex and time-consuming.

With multi-stream atomic appends and high-performance secondary indexes, KurrentDB 25.1 transforms event sourcing from a specialist technique into a mainstream development approach that any team can adopt.

The release addresses core challenges that have historically plagued event-driven architectures: maintaining consistency across multiple event streams and efficiently querying event data. Multi-stream atomic appends enable developers to maintain consistency across multiple event streams in a single operation, eliminating the distributed transaction headaches and complex coordination logic that typically slow development. Secondary indexes unlock flexible query patterns with up to 10 times faster performance and 80% reduction in storage requirements compared to previous approaches, all while preserving event sourcing integrity.

“KurrentDB 25.1 fundamentally changes what’s possible with event-driven development,” said Kirk Dunn, CEO at Kurrent. “We’ve removed the complexity barriers that kept event sourcing out of reach for most development teams. Multi-stream atomic appends eliminate the need for saga patterns and process managers just to maintain consistency across streams. Combined with secondary indexes that deliver 10x performance improvements, developers can now build event-driven applications with the same ease as traditional database applications, but with all the benefits of complete historical context and event sourcing.”

Multi-Stream Atomic Appends: KurrentDB excels at organizing events into many fine-grained streams – individual streams for every customer, product, order or business entity. Previously, events could only be written atomically to one stream at a time, forcing developers to implement complex coordination logic when changes needed to span multiple streams. KurrentDB 25.1 introduces support for appending events to multiple different streams within a single atomic write operation. This capability solves a fundamental challenge in distributed systems: maintaining data integrity across related entities without the overhead of distributed transactions. All events are committed together or not at all, eliminating the need for complex saga, process manager or workflow patterns. The feature supports optimistic concurrency control, allowing developers to specify expected versions for each stream involved in the operation. Multi-stream appends also enable faster ingress for streaming data, reducing the number of requests required to batch events to different streams in one operation. Utilizing multi-stream atomic appends requires upgrading to the latest KurrentDB client libraries.

Secondary Indexes: Secondary indexes enable developers to query event data using flexible access patterns beyond traditional stream-based retrieval. KurrentDB 25.1’s release supports category and event type indexes, providing efficient replacement for link-based projections that previously created uncontrollable data growth.

Performance and efficiency improvements include:

- Up to 10 times faster index rebuilds compared to previous projection-based approaches

- Up to 10 times faster read performance for category and event type queries

- Up to 80% reduction in disk space required for indexes

- Reduced write amplification with no impact on primary database size

- Support for both read and subscription operations

Secondary indexes power new capabilities in the embedded web UI, including database statistics and SQL query functionality that give enterprise customers exceptional visibility into and control over their event data.

After running KurrentDB in production for extended periods, enterprises can lose track of what is stored in the database. KurrentDB 25.1 addresses this with comprehensive database statistics that answer critical operational questions:

- How many events are in my database, spread out over how many streams?

- What event types are used in which stream categories?

- What are my longest streams?

- What event types are no longer in use?

The new Stats page in the embedded web UI displays total stream counts, total event counts, category and event type breakdowns with respective event counts, number of streams per category, longest streams and presence of explicit transactions in the database.

The new Query page enables users to explore and analyze event data using familiar SQL syntax without needing to write JavaScript or use the projections API for ad-hoc queries. SQL queries can find streams based on name masks, count events across streams, filter by event type or category and limit queries by event timestamp. These queries leverage secondary indexes for fast performance. Users can also execute queries with predicates on event data and metadata for deeper analysis.

KurrentDB 25.1 includes additional operational improvements designed for production deployments at scale:

- OpenTelemetry log export: Centralized log management through integration with Datadog, New Relic, Splunk, Grafana Cloud and other OTLP-compatible observability platforms. This extends KurrentDB's existing metrics export capabilities, allowing organizations to consolidate logs and metrics in a single backend.

- Apache Pulsar sink connector: A new enterprise connector enables users to stream events from KurrentDB to Pulsar topics, expanding the range of supported messaging systems and allowing greater flexibility in integrating KurrentDB with existing data pipelines and streaming architectures.

- Kubernetes Operator evolution: Advanced support for zero-downtime rolling upgrades, node resizing without downtime, custom pod labels, read-only replica pods, standalone read-only replicas in different clusters connecting to remote KurrentDB clusters, scheduled backups using volume snapshots with retention policies and log export to OpenTelemetry-compatible APM tools.

- Enhanced projection metrics: New metrics track state size, serialization duration and execution duration, helping developers identify performance bottlenecks and prevent projections from reaching limits before they fail.

- Performance optimizations: Server Garbage Collection Mode improvements, optimized StreamInfoCacheCapacity defaults (now 100,000 instead of dynamic sizing) that reduce memory footprint and GC impact, and faster startup times when using archived chunks deliver measurable performance gains across read/write operations, particularly benefiting high-throughput production deployments.

Additional features include Windows Service support, log record properties (headers) and various performance improvements in the Connectors and Archive features.

KurrentDB 25.1 is available immediately for download. Multi-stream atomic appends require upgraded client libraries that support the new functionality. Enterprise features including database statistics, SQL queries, OTLP log export and Apache Pulsar connector require a commercial license. The Kubernetes Operator 1.4 is available as a separate download.