Release Notes - What's new in Kyvos 2025.3.x
Here’s a list of Kyvos versions for Kyvos 2025.3.x releases:
Features and Enhancements for 2025.3.2
The Kyvos 2025.3.2 release brings the following new features and enhancements.
Apache Iceberg support for Spark-based semantic models on AWS
Kyvos now supports Apache Iceberg in Spark-based semantic model processing on AWS EMR version 6.15.0. This allows users to enable Iceberg support directly through Kyvos Manager.
Once enabled, users can:
Create datasets from Iceberg tables.
Define relationships between these datasets.
Build and process semantic models using Iceberg-based data.
This integration allows Kyvos to leverage Spark’s high-performance Query Engine with Iceberg’s advanced data capabilities, ensuring better performance, improved data governance, and greater scalability for enterprise analytics.
API Enhancements
Update Dataset REST API Enhancement: Users can now directly update a dataset using the dataset ID and name instead of making XML changes.
Deep Copy Semantic Model Rest API Enhancement: The API now allows the users to copy the semantic model in existing folder. The updated API also provides details of copied entities in the response.
To know more, refer to the Kyvos REST APIs reference guide for 2025.3.2
Features and Enhancements for 2025.3.1
The Kyvos 2025.3.1 release brings the following new features, enhancements, and bug fixes.
Enhanced XIRR calculations and filtering on Spark architecture: This release has enhanced support for XIRR (Extended Internal Rate of Return) calculations, featuring the following enhancements:
Ability to visualize XIRR using time series data.
Support for date dimensions at the year, month, week, or day level.
Filtering by specific date ranges is supported.
Support for scenarios where the initial net cash flow value is zero.
Performance optimizations ensure that XIRR calculations remain accurate and efficient, even with large datasets.
Improvements to Query Performance on Spark Architecture for Session Set and Response Subset Queries: This release features optimizations designed to enhance query execution time and accuracy in Spark-based deployments. These enhancements minimize unnecessary processing and more effectively manage calculated measures and query syntax. Extensive testing has been conducted in various scenarios, both with and without session sets, consistently demonstrating superior performance in all cases.
Improved query performance by eliminating semantic model parsing during execution: With this release, Kyvos has streamlined the query execution process by removing the need for multiple semantic model parsing operations. This enhancement guarantees consistent query performance, regardless of the number of calculated measures in the semantic model. Users can now add as many calculated measures as they require, ensuring optimal query performance without any negative impact.
Conditional Column Masking on Measures: Kyvos now supports conditional data masking on measures for No-Spark architecture, ensuring sensitive data remains protected and only accessible to authorized users. With this enhancement, users can apply measure masking when the conditional column is not taken into view.
Masking can be applied to base measure only.
Masked value can only be set to 0 or NULL.
Enhanced Row-Level Security (RLS) for metadata queries on Spark architecture: With this release, the metadata query of a column respect all the RLS filters applied on any column of same dimension i.e. if RLS filters is applied on any attribute it would be applicable on all attributes and all levels of all hierarchies of same dimension.
Semantic Model Enhancements
Semantic model cache purge time configurations: Users can now configure the MOLAP semantic model cache purge time at the semantic model level by adding the kyvos.cube.cache.cleanup.frequency property. This enhancement provides flexibility and control, allowing users to set specific cache durations for each semantic model and ensuring consistent query performance.
Purging of all querying-related data when semantic model is deleted: With this release, all querying data for a semantic model is purged from the Kyvos repository as soon as the semantic model is deleted. This ensures optimal space utilization and
Elimination of duplicate query export: With this release, duplicate queries will no longer be included when exporting query details from semantic models. This ensures that the size of the exported query details remains manageable over time, preventing potential failures of the export REST API due to size limits.
Usability Enhancements
This release comes with the following Kyvos usability enhancements:
Support for moving multiple measures into a measure group: Users can now move multiple measures into a measure group, ensuring data integrity while maintaining acceptable performance—even when working with a large number of measures.
Tooltips on Crosstabs: Tooltips are now supported on crosstabs, providing users with quick access to additional context and insights by simply hovering over data points. By adding measures to this shelf, any measure placed there will appear in the tooltip for crosstab visualizations. This enhancement improves data readability and makes complex reports easier to interpret.
Streamlined chart series overlap: To improve readability in charts with multiple overlapping series, the width of each series is automatically reduced by 10% for each additional measure. This ensures that all data points remain visible and easily distinguishable, enabling users to compare values clearly without visual clutter.
New Properties
This release comes with the following new property:
kyvos.process.compute.jobs.parallelism: Specifies the number of parallel jobs submitted to the compute cluster for processing each semantic model.
Features and Enhancements for 2025.3
The Kyvos 2025.3 release brings the following new features, enhancements, and bug fixes.
Here is the list of features and updates included in the Kyvos 2025.3 release.
Kyvos Modern Architecture (No-Spark Ingestion) Enhancements
This release includes the following no-Spark enhancements:
Support of raw data (Relational Multidimensional model) queries through SQL interface via metadata processing on Databricks SQL Warehouse.
Enhancements in support of filtering on calculated members in Kyvos Viz. and SQL.
Conversational Analytics Using Gen AI-Powered Kyvos Dialogs Enhancements
This release comes with the following Conversational Analytics enhancements:
Refined user experience with Kyvos Dialogs: With this release, Kyvos has enhanced the user experience in Kyvos Dialogs by introducing a more intuitive, responsive, and streamlined interface, enabling users to interact with data more effectively. For further details, see Working with Kyvos Dialogs for Conversational Analytics.
NOTE: The naming convention of Kyvos Dialogues has been changed to Kyvos Dialogs.
Upgraded Apache Tomcat Version
Kyvos has now upgraded the Apache Tomcat version for Kyvos and Kyvos Manager to 10.1.36
Role-based access to Query Playground
Kyvos has introduced an extra layer of security and customization for the Query Playground feature. This feature will now be visible only to users with the ‘Semantic Model Execute’ privilege and the new ‘Query Playground’ privilege, ensuring that only authorized users can access and use it.
Cloud Support Enhancements
This release comes with the following enhancements for AWS, Azure, and GCP clouds.
Dedicated compute cluster support
For AWS and GCP automated and wizard-based deployments: Kyvos now supports processing data on a dedicated compute cluster. Additionally, users can also switch from Kubernetes, shared Query Engine, or External Compute (EMR or Dataproc) to a dedicated compute cluster from Kyvos Manager.
For Azure wizard-based deployments: Kyvos now supports processing data on a dedicated compute cluster. Additionally, users can also switch from Kubernetes, shared Query Engine, or External Compute (Databricks) to a dedicated compute cluster from Kyvos Manager.
Support for the latest Kubernetes version on AWS (EKS), Azure (AKS), and GCP (GKE) for fresh deployments and cluster upgrades.
For EKS and GKE, Kubernetes version 1.32 is supported.
For AKS, Kubernetes versions 1.30 and 1.31 are supported.