Releasesby GlyphFlink Table API 2.3-6 adds AI functions, changelog functions, artifact creation
Java plugin 2.3-6 and Python package 2.3.3 on Flink 2.3.0 add AI_SENTIMENT, AI_DETECT_PII, AI_FORECAST and AI_DETECT_ANOMALIES, FROM_CHANGELOG/TO_CHANGELOG support, and ConfluentTools.createArtifact(). Python Table API remains in Open Preview.
Also shipped Sep 11, 2026Confluent in September 2026 →
Automations in Snowflake CoWork (General availability)
Automations in Snowflake CoWork are now generally available. An automation turns a one-time report into a recurring one: Snowflake CoWork re-runs your question with fresh data on the schedule you choose and emails you the results, including a summary and key metrics, along with a link that opens the full report so you can ask follow-up questions.
Sharing foreign Iceberg tables with OpenSharing is now generally available
Sharing and reading foreign Iceberg tables with OpenSharing is now generally available. Providers can share Iceberg tables federated from foreign Iceberg catalogs using Lakehouse Federation, including to recipients using external Iceberg clients.
Sharing foreign schemas and tables with OpenSharing is now generally available
Sharing and reading foreign schemas and foreign tables with OpenSharing is now generally available. Providers can share foreign schemas and tables federated from external data sources using Lakehouse Federation, without copying data into Databricks.
Session restore for serverless jobs is in Public Preview
Restore the Python variables and Spark session from a serverless compute job run into a new interactive notebook, so you can debug a failure or explore results without rerunning the entire job. Databricks captures the snapshot automatically for eligible runs. See Session restore for serverless jobs .
Second generation Openflow deployments and runtimes now available on GCP
Second generation (gen 2) Openflow deployment and runtime SQL objects are now available on Google Cloud Platform (GCP), in addition to AWS and Azure. Previously, gen 2 deployment and runtime SQL objects were generally available only on AWS and Azure.
Add VS Code Agents to Copilot usage metrics
GitHub Copilot usage metrics reports now include generally available metrics for activity in the dedicated VS Code Agents window, helping you measure adoption and engagement across enterprises and organizations.
More from Confluent CloudEvery Confluent Cloud release →
Oracle XStream CDC Source no longer requires SELECT_CATALOG_ROLE
The connector no longer requires granting SELECT_CATALOG_ROLE to the connect user; you can instead grant SELECT on an explicit list of catalog views.
Oracle XStream CDC Source supports Active Data Guard standby as source
For cascading downstream capture, the source connection can point at the Active Data Guard standby instead of the primary, offloading the initial snapshot and schema reads.
ksqlDB v8.1.0-41 bounds pull-query concurrency per cluster
Coordinator and router thread pools each accept up to 100 pending requests; requests beyond that are rejected with HTTP 429. Queries are cancelled server-side if the client disconnects.
ksqlDB validates property overrides against an allowlist
Server property overrides supplied through queries, such as SET statements, are now validated against an allowlist of approved properties by default, replacing the previous denylist. Overrides outside the allowlist are rejected.
Logical cluster ID OAuth extension no longer required for Kafka clients
The logicalCluster OAuth extension is no longer required for clients connecting to Dedicated, Freight, and Enterprise clusters with an lkc-based bootstrap URL, on any cloud or networking type.
Advanced Security add-on in Limited Availability
Organization-level add-on bundling SCIM, self-managed encryption keys (BYOK), and higher limits for security resources such as identity pools and RBAC role bindings. Billed as a percentage uplift on eligible infrastructure charges.
SCIM user provisioning generally available
SCIM user provisioning is GA with the Advanced Security add-on. Your IdP becomes the single source of truth for SSO user accounts, automating provisioning and de-provisioning.
Flink materialized tables can adopt an existing Kafka topic
A CREATE OR ALTER MATERIALIZED TABLE statement can point at an existing table, and Flink adopts it in place, keeping the topic, its data, and its consumers. Previously a materialized table always created a new topic.