Arenadata Streaming 4.0 adds Apache Iceberg V3 and ADB 7 compatibility

On 2 July 2026, Arenadata announced Arenadata Streaming 4.0.0.b1. The release updates its streaming pipelines: Iceberg Sink Connector supports the Apache Iceberg V3 specification, NiFi ADB Connector is compatible with Arenadata DB 7, and VictoriaMetrics-based monitoring is available through Arenadata Monitoring.

What the primary source confirms

The Iceberg V3 implementation covers VARIANT and UNKNOWN types, nanosecond timestamps, deletion vectors and metadata encryption keys. These capabilities expand the lakehouse table model, while requiring compatibility checks for every engine that reads the tables.

NiFi ADB Connector adds the GetGreengageRecord processor. It extracts data from Greengage DB through gpfdist and supports incremental reads of newly available ranges. The vendor says this can reduce database and NiFi load in near-real-time scenarios.

The release also provides:

• out-of-the-box NiFi ADB Connector compatibility with ADB 7;
• ADS installation through ADCM Wizard with preconfigured SSL, Kerberos and Ranger options;
• VictoriaMetrics-based monitoring in Arenadata Monitoring;
• updated Kafka and NiFi services, plus Debezium and Iceberg connectors;
• Astra Linux 1.8 support.

Why Iceberg V3 matters to streaming

A streaming lakehouse pipeline manages more than write speed. It must handle row changes, schema evolution, metadata quality and reader compatibility. Deletion vectors can represent deleted rows without immediately rewriting every data file. Nanosecond timestamps matter when coarser precision would lose event ordering.

Connector support alone does not make the whole analytics estate V3-ready. Tables may be read by different Spark, Trino and Iceberg-library versions. Inventory every reader and writer before enabling new format capabilities.

Pilot checklist

1. Confirm that every analytics engine can read the target Iceberg V3 tables.
2. Test updates and deletions after duplicate event delivery.
3. Verify timestamp precision throughout the pipeline.
4. Prepare rollback for an older consumer that cannot read a new table.
5. Measure metadata growth and maintenance operations.
6. Review permissions for NiFi, Kafka Connect, the catalog and object storage.
7. Test GetGreengageRecord after interruption, range replay and schema changes.
8. Confirm which VictoriaMetrics dashboards and alerts appear after installation.

Use production-like record and partition-size distributions in the load test. An average throughput figure does not expose checkpoint delays, small-file accumulation or recovery after a consumer outage.

Remaining limitations

The official announcement describes release capabilities but provides no independent benchmark, maximum operating envelope or migration cost for existing tables. The version carries a b1 designation; teams should confirm support conditions and production suitability in the documentation and contract.

Check exact driver versions, authentication schemes and network policies in addition to ADB 7 compatibility. A preconfigured Wizard can speed deployment, but it does not replace threat modelling, certificate rotation and operational tests.

Practical conclusion

The release is relevant to teams connecting Kafka and NiFi to a lakehouse and Arenadata DB. Iceberg V3 functions and incremental extraction can reduce the need for custom workarounds. A sensible pilot starts with one non-critical topic, a separate V3 table and a reversible migration plan.

A production design still needs an integration map, data ownership, pipeline SLOs and source-to-table observability. See the guide to API integration architecture and infrastructure observability.

Source: Arenadata’s official announcement dated 2 July 2026.