March 02, 2026

New features

Latest NiFi version

Flow Deployments and Test Sessions now support the latest Apache NiFi 1.28.1 and NiFi 2.6.0 releases.

Multiple flows on the same deployment

Deployments and flows have been decoupled in Cloudera Data Flow and now it supports adding multiple flows to the same NiFi deployment. This allows for optimizing cloud infrastructure costs and resources associated with deployments.

Provenance events in Flow Designer

Flow Designer now provides a tool to observe data provenance events in running test sessions. It provides a searchable historical record of every data object (FlowFile) as it moves through your flow. It can help you debug and audit issues with your flows by understanding the changes (events) that occurred to data as it was being processed by each processor.

Platform updates

New Kubernetes version support

Cloudera Data Flow now supports EKS/AKS 1.33.

Changes and improvements

Reduced Flow Designer CPU consumption

This release of Cloudera Data Flow introduces an improvement that reduces CPU consumption of Flow Designer.

New test session states allow for better control of cloud costs and test session lifecycle management

Test sessions now have states identical to those of deployments (active,suspended, terminated). Suspending a test session tears down associated cloud resources while preserving the assets associated with the draft for which the test session was initiated, making them available for future use. Terminating a test session removes all associated assets and related cloud resources.

Projects enforced on environment level

Association of resources to a project can now be enforced on the environment level. If this feature is enabled for an environment, users cannot create resources with 'Unassigned' status. You can switch this feature on while enabling an environment for Cloudera Data Flow.

Project-level usage tracking

With a new 'Project' tag added to Cloudera Data Flow usage events, it is now possible to run reports for Cloudera Data Flow usage with project-level granularity on the Cloudera Management Console.

NiFi End of Support (EoS) policy enforced during service upgrades

When running an environment upgrade with the 'Preserve NiFi versions of deployments and test sessions' selected to keep operability of your existing resources, the upgrade wizard checks if the NiFi runtime versions of those resources will be compatible with the upgraded Cloudera Data Flow service. If the result is negative, the upgrade is blocked until you either upgrade the offending resources to the minimum supported NiFi version, let the wizard upgrade them to the latest available version as part of the standard upgrade process, or delete them.

Updated ReadyFlows

• The RAG Query Pinecone ReadyFlow has been updated so the "Bedrock Model Name" no longer defaults to a specific model and instead instructs users to choose an active, non‑deprecated model.
• In the Airtable to S3/ADLS ReadyFlow the 'Airtable API Key' configuration parameter has been replaced with 'Airtable Personal Access Token'.

Removed ReadyFlows

• ADLS to Chroma DB
• S3 to Chroma DB
• S3 to S3 Avro with S3 Notifications
• Slack to Chroma DB

Fixed issues

Lost Flow during zero leader scale up

If a NiFi cluster lost all viable leaders during scale up (usually a single node cluster scaling to two nodes), the new Node would begin waiting for leader election, eventually time out, and declare itself the leader with an empty Flow definition. When the old leader finally recovered, it would take the empty Flow of the leader, assuming it had no data in queues. The CFM Operator will now temporarily revert the scale up if there are no viable leaders at any point in the scale up process.