Additional Items in v26

The following smaller improvements complement the major v26 platform changes.

FHIR devices and transformation

v26 adds purpose-built devices for converting HL7 v2 messages to FHIR R4, validating FHIR resources with the Firely FHIR .NET packages, and delivering resources and bundles to FHIR REST endpoints. The Liquid engine also supports custom transformations whose source or target is not FHIR.

See FHIR Devices and Transformation for the device choices, validation modes, and end-to-end flow.

Configuration and unattended setup

v26 adds a cross-platform terminal wizard and a command-oriented configuration CLI for Windows, Linux, containers, and automated deployments. The CLI supports guided setup, non-interactive commands, and JSON output for scripts and deployment tooling. Connexion install configuration and Remote Agent configuration are now file-backed rather than dependent on the Windows registry, and supported installation paths can automatically install PostgreSQL instead of the former LocalDB bootstrap option.

See Linux Support and PostgreSQL Support for the related platform and database changes.

Remote Agent operational improvements

Remote Agent and Gateway operations include several smaller usability and reliability improvements. Gateway operators can select multiple Remote Agents for supported batch operations. Package and updater handling is more resilient, registration failures produce clearer errors, each Remote Agent publishes its own host-information metrics, and Remote Agents can report low disk space for their managed MongoDB instance.

The Remote Agent service also benefits from the broader .NET 10 and Linux platform support and the v26 observability model.

Generic database providers and code generation

Generic database features no longer depend on machine-wide ADO.NET provider registration. SQL Server and PostgreSQL providers are built in, while additional databases can be enabled through Connexion provider packages. Database Poller and Custom Code surfaces obtain available providers from this Connexion-managed catalog.

The Custom Code data-context generator has also moved from the legacy Entity Framework model to Linq2Db source generation. Generated contexts are supported for SQL Server, PostgreSQL, MySQL, Oracle, SQLite, and Firebird and compile against the v26 runtime.

See Plugin Developers: What You Need to Know for v26 for the broader custom-code and plugin migration story.

Optional Kafka message ordering

Kafka Writer devices can enable PreserveMessageOrder when messages from a channel must remain ordered through retries. When enabled, the writer uses an idempotent producer and sends the channel's messages through a single Kafka partition. The default remains unordered so that all available partitions can be used for greater throughput.

Performance and efficiency improvements

v26 includes a collection of lower-level efficiency improvements in addition to the redesigned queues and large-message architecture. These include runtime-role-specific garbage-collection tuning, pooled compression buffers, reduced allocations in HL7 serialization, incremental statistics rollups, and streaming export paths. Together they reduce avoidable memory use and processing overhead without changing the familiar channel and device model.

See SQL Server Queue V2 and Large Message / Attachment Support for the larger performance-related architectural changes.