These are the Treasure Data features and improvements released in June 2025.
Smart Schedule — Time Range Option for CJO Activations
We’ve enhanced CJO scheduling with Smart Schedule, introducing a new Time Range option.
Marketers can now specify a flexible window (e.g., 2:00–3:00PM) for activations to run, without worrying about underlying workflow or segment refresh dependencies.
This ensures campaigns are sent at the optimal time based on when the latest data is ready, reducing manual coordination and complexity are parent segment refresh times.
Enjoy greater control and confidence in your activation timing!
Learn More: Treasure Data Product Documentation
Text Knowledge Base: Enhanced Agent Prompting in AI Framework
Lets users define custom multiline text content as agent-accessible knowledge bases which agents can reference directly.
Enables agents to easily access and retrieve instructions, templates, and frequently used phrases without database queries, improving workflow efficiency for prompt engineering and agent tasks.
- Eliminates the need for database queries when only text is required, improving efficiency and expanding potential AI use cases.
Removes previous limitations of supporting only database sources, laying the groundwork for broader knowledge base support and enhancing flexibility for more diverse AI use cases.
Learn more: Treasure Data Product Documentation
AI Agent Foundry Events Now Included in Premium Audit Log
Premium audit logs now capture create, modify, and delete actions for LLM Form Interfaces and LLM Text Resources.
Enhances transparency and auditability for customers leveraging AI Agent Foundry solutions.
Helps organizations monitor and track critical updates to AI-driven content management features.
Custom Script Tasks - Launch Time Improvement
Custom Script Workflow tasks now start up to 3x faster with the integration of Seekable OCI (SOCI) technology.
Launch times are reduced from approximately 60 seconds to just 20 seconds, streamlining data engineering workflows and accelerating workflow execution.
Improves productivity and also reduces operational infrastructure costs.
td_wait > Operator - Now Supports Timeout Option
The Treasure Workflow td_wait> operator now supports a new wait_timeout parameter, allowing users to specify a maximum wait time for data arrival events.
This enhancement helps prevent workflows from waiting indefinitely, improving reliability and aligning td_wait> with similar operators that already support a timeout option.
The wait_timeout parameter was intentionally chosen to ensure backward compatibility with existing workflow configurations; no changes or action are required for existing users.
Learn More: Treasure Data Product Documentation
Custom Script Docker Image Versioning Policy Update
Security fix versions (e.g., "3.9.1-3") are now the default for Custom Script Docker images, enabling urgent security patches and upgrades with reduced risk of breaking changes.
Major and minor version images ("3.9", "3.9.1") are discontinued; users should always reference the complete security fix version for consistency and reliability.
Image release cycles are now flexible. Images will be updated as required, without a fixed schedule, to better address evolving needs.
Learn More: Treasure Data Product Documentation
Custom Script Docker Image Update – Python 3.9.22-td1
New Docker image released for Custom Script workflows, now using the latest pytd library (3.9.22-td1) to ensure up-to-date dependencies.
Addresses security vulnerabilities and keeps dependencies up-to-date for improved safety and compatibility.
Update is available across all Treasure Data regions for immediate use.
Learn More: Treasure Data Product Documentation
Treasure Workflow Notification Enhancements
Expanded notification coverage for workflow failures, including scenarios that previously produced no alerts (like permission issues or workflows with too many tasks).
Enhanced architecture now ensures notifications are sent even if a workflow never starts, eliminating silent cancellations.
Users can more reliably monitor workflow execution and quickly respond to errors for improved operational transparency.
Learn More: Treasure Data Product Documentation
Real-Time Parent Segment Creation Improvement
Parent segment owners can now create and launch real-time configurations and journeys without requiring Treasure Data admin privileges.
This change streamlines workflow, reduces bottlenecks, and enables broader team collaboration for segment management.
Teams gain greater operational flexibility and efficiency when orchestrating real-time customer journeys.
Learn More: Treasure Data Product Documentation
HTTP > Operator Audit Logging Enhancement
All executions of the http> operator in workflows are now logged in the premium audit log to increase compliance and traceability.
Logged details include the workflow project ID, project name, session and attempt IDs, task path, and the called HTTP endpoint (excluding query strings), all recorded in the event_detail column as JSON.
This change is designed to help customers, especially those in regulated industries like finance, monitor and audit data transfers to external endpoints, enhancing security and governance.
Note: There is no change to the audit log schema, but users parsing event_detail with a fixed schema should verify compatibility.
LinkedIn Audience Connector – Token Lifetime Display
The Authentication page for the LinkedIn Audience Connector now shows the token lifetime, allowing users easily see when their LinkedIn authorization will expire.
Users are proactively informed when to reauthenticate, helping prevent unexpected disruptions to data activation.
The system leverages LinkedIn’s Token Introspection API to provide accurate, real-time token validity every time the page is visited.
Learn More: Treasure Data Product Documentation
Behavior Activation in CJO Activation Templates
Marketers can now configure and lock behavior data for export directly within Customer Journey Orchestration (CJO) activation templates, providing consistency with batch segment capabilities.
Only published templates with matching behavior configurations are selectable for activations, ensuring data accuracy and relevance for each journey.
Error handling, field validation, and template management are improved, delivering a straightforward and reliable experience for creating and editing activations.
This update empowers users to build more sophisticated, scalable customer journeys by reusing behavioral logic, enforcing template consistency, and reducing activation errors.
Behavior Activation Limitations Removed
Advanced segmentation now offers greater flexibility. Previously, limitations existed where users could not reuse behavior tables across multiple rule sets, apply "Match Logic ANY" for flexible audience definitions, or handle missing behavior data using null values.
These restrictions have been removed, allowing for truly dynamic segments and fully leveraging behavior-driven personalization in journey activations.
Rank and Limit for Advanced Segmentation (Beta)
Marketers can now rank customer profiles by key behaviors (e.g., purchase value, email activity) and restrict segments to a fixed number or percentage of top profiles—directly in the Segment Rule Builder, without writing SQL.
Supported in both Batch Segment Builder and Journey Rule Builder; works with multiple behavior table rules in a single segment.
Sort profile results in ascending/descending order and select the top N or top X% of profiles based on chosen criteria.
Meta Offline Conversion API Connector – Deprecation Notice
The Meta Offline Conversion API connector will be officially retired following Meta’s own deprecation of the Offline CAPI API in May 2025.
Customers should migrate to the Meta Online (unified) Conversion API to continue sending conversion data.
The Offline Conversion API connector will be removed from the Treasure Data catalog; support and documentation updates are available for guidance.
Data Connector: BigQuery Import Connector V2
Import table data (including partition tables) and query results from Google BigQuery directly into Treasure Data.
Enhanced support for BigQuery data types: NUMERIC, BIG NUMERIC, DATE, TIME, DATETIME, BYTES, and GEOGRAPHY.
Incremental data ingestion now available based on a single integer or timestamp column, streamlining ongoing data synchronization.
Materialized Views now supported via query imports; leverage BigQuery User Defined Functions (UDF) during import.
Learn More: Treasure Data Product Documentation
LinkedIn Ads Conversions API Export Integration: Enhanced Dropdown Experience
The LinkedIn Ads Conversions API Export Integration has been enhanced to improve user experience within Audience Studio by dynamically populating the Ad Account and Conversion Rule dropdown menus.
Ad Account dropdown now loads dynamically when you open the configuration screen, showing only relevant LinkedIn Ad Accounts.
Conversion Rule dropdown is populated in real time based on the selected Ad Account, streamlining and simplifying the setup process.
Improved validation and usability for editing and scrolling through dropdown options with a smoother, more efficient workflow.
Learn More: Treasure Data Product Documentation
LiveConnect: Redshift – Easy Sync API
Easy Sync is an API-driven workflow designed to simplify and accelerate the process of copying multiple tables from Redshift into Treasure Data.
Enables seamless, automated copying of multiple tables from Redshift into Treasure Data via an API, eliminating manual and complex workflows.
Delivers high performance and scalability with advanced parallelism, supporting large-scale data migrations and overcoming previous concurrency and throughput limitations.
Designed for enterprise data engineers seeking efficient, reliable, and scalable data warehouse integrations who frequently move large volumes of structured data between Redshift and Treasure Data.
Learn More: Treasure Data Product Documentation
LiveConnect: BigQuery - Easy Sync
Simplifies complex, multi-stage migrations into an efficient, first-class experience for high-volume pipelines.
Empowers organizations to move structured data quickly for marketing and customer experience initiatives.
Effortlessly copy multiple tables from Google BigQuery to Treasure Data with a streamlined, API-driven workflow.
Advanced parallelism unblocks typical concurrency bottlenecks for faster, high-volume data migration.
Specifically designed for data engineers managing 100+ table transfers in large-scale enterprise environments.
Learn More: Treasure Data Product Documentation
US Console Domain Change – July 2025
Starting in early July 2025, the Treasure Data Console domain for US users will be updated, leading to an automatic logout for all users, who will need to log back in once the update is complete.
All users will be automatically logged out as part of this update.
You must sign back in after the domain change is complete.
This update supports ongoing infrastructure improvements and enhanced security for US region users.
Presto ODBC Driver End of Support (July 2025)
Treasure Data will discontinue support for the packaged Presto ODBC Driver (powered by Simba) by July 2025.
Customers are advised to migrate to alternative solutions, such as JDBC-based connections, Connector integrations, or third-party Presto/Trino ODBC drivers.
Future development and troubleshooting for the Simba-based ODBC driver will cease after the deprecation date.
This deprecation aims to streamline support and encourage adoption of modern connection methods.
Deprecation of "Partial Delete Job" Feature – US Region
The “Partial Delete Job” feature, including the td table:partial_delete command, will be deprecated in US regions as of July 2025.
All users should migrate to the Trino DELETE statement, which provides greater flexibility with custom deletion conditions.
The “Partial Delete” operator will also be removed from Treasure Workflow commands.
This aligns the US region with changes already implemented in Tokyo, EU, and APAC regions.
Trino Label Update: Console and API Changes
The query engine "Presto" is now renamed to "Trino" in the Treasure Data Console and APIs.
Users will see "Trino" supported as a parameter in API calls and across Console labels, aligning with current industry standards and engine maintenance responsibilities.
This change helps reduce confusion between Presto (maintained by Facebook) and Trino (maintained by Starburst), ensuring clearer documentation and search results.
No action is required for most users; however, those using API parameters or searching for documentation should now reference "Trino."
Related libraries (CLI, Workflow clients) are also being updated to reflect this naming change.
Advanced Workflow Management: Session Pool (Private Beta)
Users can prioritize and efficiently manage large volumes of workflows through configurable session pools.
Helps ensure that critical workflows run first and system resources are allocated based on business priorities, even under heavy operational loads.
Enables prioritization, allowing important workflows (such as emergency campaigns) to execute immediately by assigning them higher priority within session pools.
Improves overall system efficiency by dynamically allocating workflow concurrency limits and maximizing resource utilization.
Provides administrators with flexible, rule-based grouping and concurrency management via an intuitive API, supporting up to 10,000 workflows without performance degradation.