Better workflows, better visibility, and new power tools for AI development.
This release brings major enhancements across Domino Apps, Governance, and Data - with smoother workflows, autoscaling infrastructure, and audit-ready controls that help you scale up without slowing down.
Additional component updates are available in ddlctl and fleetcommand-agent Release Notes.
-
Apps autoscale automatically and can act with viewer permissions for secure, flexible deployments
-
Cut compute costs with AWS Spot Instances across Jobs, Workspaces, Apps, and Endpoints
-
GxP-ready compliance with workspace audit trails, Posit Package Manager, and governance automation
-
Register and deploy LLMs with OpenAI-compatible APIs for production agentic systems
-
Real-time governance autosave, policy gates, and monitoring integration strengthen controls
-
End-to-end agentic system development from instrumentation through production deployment
Apps are now easier to scale, discover, and secure with smarter defaults and granular control when you need it.
Autoscaling for App Deployments
Keep your apps responsive under variable load with horizontal pod autoscaling:
-
Automatic scaling based on CPU and memory usage
-
Default autoscaling behavior for faster setup
-
Optimized resource usage across the board
Improved App Access Management
Control app access more precisely with redesigned sharing controls and automatic notifications:
-
New Access & sharing section with adaptive controls
-
Clearer visibility settings for project-wide and public apps
-
Email and in-platform notifications when users are granted access
Allow Apps to Act for Viewers in Domino
Apps can now act with the viewer’s Domino identity instead of the publisher’s:
-
Viewers grant permission when they open the App, allowing it to access data and perform other Domino actions on their behalf
-
Consent is time-limited and auditable
-
Only SysAdmin or CloudAdmin roles can publish Apps with this feature enabled
Take the guesswork out of agentic systems development with end-to-end tracing, structured evaluations, and seamless production deployment.
LLM & AI System Monitoring
Monitor agentic systems and LLM applications with full trace visibility:
-
Capture every LLM call, agent interaction, and prompt chain as traces that show inputs, outputs, and downstream tool calls.
-
Attach evaluations to traces to measure quality, accuracy, or safety
-
Compare performance across experiments to understand what’s working and what’s not
Production Logging & Traces
Capture every call your agentic system makes in production automatically:
-
Works with MLflow auto-tracing integrations (LangChain, OpenAI, Pydantic AI)
-
Supports standalone functions without specific frameworks
-
Logs all traces, parameters, metrics, and artifacts in experiment runs
Deploy AI Systems to Production
Deploy your agentic systems with infrastructure optimized for agent workloads:
-
Use the same proven infrastructure as Domino Apps
-
Instrumentation carries over automatically from experimentation
-
Capture production traces from real user interactions without additional setup
Enhanced policy management, automated monitoring integration, and streamlined workflows make governance easier to implement and maintain.
Additional Governance Improvements
Streamline governance workflows with these enhancements:
-
Bundle History: Track every change to your bundles with the new Bundle History tab showing what changed, when, and by whom.
-
Governance Monitoring Integration: Respond automatically to Domino Model Monitoring alerts by creating findings, assigning severity levels, and notifying owners when drift or data quality issues trigger.
-
Sequential Workflows with Required Fields: Create ordered approval flows where each stage must be fully completed before the next begins, with required evidence fields enforced at every step.
-
Multiple Policies per Bundle: Apply multiple policies to a single bundle to reduce duplication and streamline reviews.
-
Archive Unused Policies: Archive policies from the Policies screen while keeping them visible with filters.
-
Enhanced PDF Bundle Reports: View attachments in a dedicated appendix section for easier review.
-
Flexible Approval Assignment: Assign different approvers to individual bundles under the same policy for context-specific reviews.
-
Improved Bundle Management: Assign individuals to specific stages, manage multiple bundles simultaneously, and customize terminology to match industry standards.
-
Attach Snapshots and Flow Artifacts: Link NetApp Volume snapshots and Flow artifacts directly to bundles for complete traceability.
Move data seamlessly between tools, preview more file types directly in Domino, and connect to data sources with certificate-based authentication.
Flow Integration with NetApp Volumes
Connect Flow artifacts and NetApp Volumes for seamless data movement across executions.
-
Export Flow artifacts to NetApp Volume snapshots via UI or Flyte Python SDK
-
Mount snapshots directly in Flows at runtime for read-optimized data access
-
Make Flow outputs immediately available to downstream tools like Apps
Snowflake Single-Factor Authentication Deprecation
Prepare for Snowflake’s deprecation of single-factor authentication:
-
New Snowflake datasources require multi-factor authentication
-
Existing datasources display warnings to encourage updating authentication methods
-
Update now to avoid disruption when Snowflake completes the deprecation
Track model performance and system activity with enhanced monitoring dashboards and improved audit capabilities.
Model Endpoint Monitoring Updates
Monitor model health and performance with enhanced real-time visibility:
-
View real-time model metrics directly in Domino
-
Access advanced dashboards in Grafana for deeper analysis
-
Debug issues faster and optimize resource usage
Workspace Audit Capabilities
Extend Domino’s audit trail into interactive workspaces for complete visibility and regulatory compliance:
-
Log all file operations (read, write, create, rename, delete) on NetApp Volumes and Datasets
-
Investigate activity quickly with a dedicated UI for searching, filtering, and exporting events
-
Customize the audit app or integrate programmatically with governance tools
-
Retain events long-term in customer-managed blob storage with automatic deduplication
-
Meet GxP and SOC 2 requirements with comprehensive, auditable records of in-workspace data access
Streamline development workflows and reduce compute costs with enhanced package management and flexible infrastructure options.
Posit Package Manager
Simplify R package management in RStudio and VS Code with GxP-ready reproducibility:
-
Search for packages across PPM snapshots
-
View which snapshot is configured in your
renv.lockfile -
Install packages with a single command
-
Ensure validation and reproducibility for regulated Life Sciences workflows
Reduce Compute Costs with Spot Instances
Cut compute costs by using discounted, interruptible AWS cloud capacity across Jobs, Workspaces, Apps, Endpoints, and Launchers:
-
Select hardware tiers marked with the Spot tag to use spot pricing automatically
-
Ideal for batch jobs, distributed training, and parallel analytics that tolerate interruptions
-
Domino falls back to on-demand infrastructure if spot capacity is unavailable
-
Interrupted jobs retry automatically on available spot capacity or on on-demand instances
-
Available on AWS with automatic capacity management
Simplify API authentication and track compute costs more accurately with these platform improvements.
Authenticate with a Personal Access Token
Generate short-lived tokens for API authentication directly from your account settings:
-
New Personal Access Token panel in account settings
-
Use tokens with the Domino API or any API endpoints
-
Secure, time-limited access without sharing credentials
-
Fixed issue where Service Account tokens contained mismatched User ID information, causing authentication failures.
-
Fixed issue where Unified Audit Trail would stop ingesting events if custom pod labels have been set for the cluster and network policies were enabled. This change requires a manual deletion of the fluent-bit daemonset if custom global pod labels were set for the cluster before the upgrade.
-
Fixed issue where one misconfigured endpoint caused the mlflow-deployments pod to crashloop and AI gateway to fail with 502 errors.
-
When annotating a Domino Flows sub workflow’s outputs as a Flows Artifact, the artifacts will now be properly exposed in the Artifacts UI for that Flow.
-
After upgrading to version 6.1.1 from 6.1.0, Workspaces running in the local data plane with SSH enabled must be restarted once to retain SSH access.
-
Flows is only supported in Domino Cloud, AWS, and Azure. A future update will add support for Google Cloud and on-premises installs.
-
Some Flyte features are not yet available in Flows, such as dynamic workflows and map_task.
-
In Azure Blob Store deployments, Projects with many files may fail to sync through the Domino CLI. To work around this issue, do not disable file locking when prompted by Domino.
-
You cannot view the latest raw file if you click View Latest Raw File. In the navigation pane, go to Files and click a file to view its details.
-
If idle timeout is configured on a deployment, Domino may not recognize pages opened in spawned tabs (e.g., via
Ctrl+Click) or outside of an iFrame as being active. Users may be unintentionally logged out despite activity in the spawned pages.
-
When uploading a large file to the Azure blob store by syncing a Workspace, you may encounter a Java Out of Memory error from Azure if the file/blob already exists. To work around this issue, use the Domino CLI to upload the file to the Project.
-
Model Monitoring data sources aren’t validated. If you enter an invalid bucket name and attempt to save, the entry will go through. However, you won’t be able to see metrics for that entry, because the name points to an invalid bucket.
-
When restarting an RStudio workspace with home directory persistence enabled, the user may see a notice stating that the previous session was abnormally terminated due to a crash. This notice is only due to how the new feature interacts with RStudio and does not result in any data loss.
-
Domino instances that make use of Azure Blob Storage may experience stalled Jobs within Projects with many large files.
-
If you attach a Git repository to a DFS Project that points to a tagged release, the tag won’t be honored when building a Domino endpoint in that Project. The build log will show an error similar to the following, and the model will be built using the default branch of your Git repository instead of the tagged branch:
Jul 05 2023 14:36:27 -0500 #10 6.481 WARN [d.r.d.GitRepoUpdater] could not parse ref: v1.3.0 checking out default branch correlationId="iA2qWrYSLQ" thread="main"
To work around this issue, use the branch name when building Domino endpoints instead of the release tag.
-
If an admin resets a user’s password, it invalidates all the user’s authentication tokens, including tokens used for long-running tasks like Jobs, Workspaces, or Apps. The user must create a new password, log back into Domino, and restart all executions. This also applies to CLI authentication; the user must re-login to their Domino CLI.
-
Data storage page fails to load when the dataplane agent in the remote data plane is scaled down to zero.
-
Nucleus Dispatcher restart hangs during Quartz Scheduler startup following a MongoDB replica set scale-down. As a workaround, manually remove the unavailable replica from the replica set using the Mongo CLI command.
-
During a 48-hour soak test, the
nginx-ingress-controllerpods exhibited a gradual increase in memory usage while handling heavy Domino Endpoint workloads. The cause is under investigation.
-
Agent monitoring dashboard retains only the current day’s data (charts have missing data and traces are gone). Historical data from previous days are lost, regardless of what time window is selected.
-
In large-scale environments, Keycloak multi-replica deployments may hit Infinispan cluster timeouts under load, resulting in login failures. Workaround is to scale Keycloak down to a single replica (or re-scale after stabilization).
-
Gen AI model registration fails on Azure and GCP because image existence checks don’t work with ACR or GCR registries. Workaround: skip the image existence check. Downside: re-registering a model will rebuild the image even if it already exists.
The following versions have been validated with Domino 6.2.0. Other versions might be compatible but are not guaranteed.
-
Kubernetes - see the Kubernetes compatibility chart
-
Ray - 2.43.0
-
Spark - 3.5.4
-
Dask - 2025.10.0
-
MPI - 4.1.4
Component and compatibility updates this release:
-
OpenShift version compatibility updated to 1.32 (OpenShift 4.19).
-
GKE version compatibility updated to 1.33.
-
AKS version compatibility updated to 1.33.
-
EKS version compatibility updated to 1.33.
-
Keycloak version upgraded to 26.3.5.
-
RabbitMQ version upgraded to 4.1.2.
