Release Notes - What's new in Kyvos 2025.4.x

Here’s a list of Kyvos versions for Kyvos 2025.4.x releases:

• Kyvos 2025.4.2- Features and enhancements

• Kyvos 2025.4.1 - Features and enhancements

• Kyvos 2025.4 - Features and enhancements

Features and Enhancements for 2025.4.2

The Kyvos 2025.4.2 release brings the following new features and enhancements.

Support for GCloud Storage Commands

With this release, Kyvos has removed the dependency on aws-cli and gsutil commands for Kyvos upgrades, patch deployments, Kyvos Reporting deployment, and Elastic Search deployment. This change aligns with Google Cloud Platform (GCP) best practices and recommendations.

Improved Credential Handling for GCP Deployments

In GCP enterprise environments, the Kyvos Manager UI now supports a more secure and streamlined method for authentication. For version-based patch deployments (Kyvos Manager or Kyvos), Kyvos Manager upgrades, and Kyvos Reporting deployments, users are now required to provide a base64-encoded service account key instead of release bucket access credentials.

Configurable Compute Nodes for Semantic Model Jobs (AWS Only)

Kyvos now supports configurable compute node allocation for semantic model data processing in dedicated compute clusters on AWS. This enhancement enables users to fine-tune performance and optimize costs by selecting the appropriate number of nodes based on their workload requirements.

Features and Enhancements for 2025.4.1

The Kyvos 2025.4.1 release brings the following new features and enhancements.

Configurable compute nodes for semantic model jobs

For Azure and GCP environments, Kyvos now allows users to select the number of compute nodes allocated for processing semantic model data in a dedicated compute cluster. This enhancement allows users to fine-tune performance and manage costs by selecting the optimal number of nodes for their data processing workloads.

Improved Reliability with Load-Based Query Scheduling

Kyvos now offers enhanced stability for query execution when using load-based scheduling. If a query is triggered while Query Engines are in the process of stopping, it will no longer fail. Instead, the query will be automatically held in a queue and executed once the engines are back online—ensuring uninterrupted insights and a smoother user experience.

Enhanced Query Resilience During Query Engine Restarts

Kyvos now ensures greater reliability for query execution, even during Query Engine restarts. If a query is triggered while the engines are stopping or already stopped, it will no longer fail. Instead, the query will be held and automatically processed once the engines restart. By default, queries are held for up to five minutes, with the option to customize this duration using the QE_STARTUP_QUERY_HOLD_TIME property.

Features and Enhancements for 2025.4

The Kyvos 2025.4 release brings the following new features and enhancements.

Here is the list of features and updates included in the Kyvos 2025.4 release.

• Features and Enhancements

• Cloud Support Enhancements

• Kyvos Viz. Enhancements

• API Enhancements

Kyvos Modern Architecture (No-Spark Ingestion) Enhancements

This release includes the following no-Spark enhancements:

• Enhanced support for filtering on calculated members in Excel.

• Kyvos now supports joining two or more raw data semantic models in Tableau. This is supported only when using Snowflake as the raw data SQL engine.

• Query performance enhancement by distribution of queries across multiple Query Engines on the cluster in a round-robin manner, such that the complete query can be served from a single Query Engine instead of reading cuboids from multiple Query Engines.

• Support of Slowly Changing Dimensions (SCD 1) in semantic model processing.

Support of global parameters for datasets

Kyvos now supports global parameters for datasets, making it easier to manage parameter values (like dbname and tablename) during environment migrations (e.g., from UAT to Production). You can set these parameters in the connection settings, and Kyvos will automatically use the right values based on the target environment. This reduces manual work, improves deployment flexibility, and simplifies environment management.

Internationalization Enhancements

As part of ongoing enhancements, this release includes internationalization support in Kyvos Manager for users’ preferred language in dynamic messages such as operation logs, cluster events, notifications, and audits.

Cloud Support Enhancements

This release comes with the following enhancements for AWS, Azure, and GCP clouds.

• Adjustable Query Engine Instance Capacity: Kyvos now lets you change the Query Engine instance capacity to better match your performance and cost needs. With this feature, users can increase or decrease the maximum Query Engine instance type to boost query performance during peak usage. This helps improve query performance while controlling infrastructure costs.

• High Availability for Compute Server (Dedicated Compute Cluster): Kyvos now supports High Availability for the Compute Server in dedicated compute clusters. This ensures continuous service and automatic recovery from unexpected downtimes.

AWS Enhancements

• Support of Shared Query Engine and dedicated compute clusters in AWS SaaS (No-Spark deployments): With this release, Kyvos supports processing semantic models using either Shared Query Engines (QE) or Dedicated Compute Clusters for No-Spark deployments in AWS SaaS environments.
  Please note that Kyvos 2025.4 onwards, Spark (EMR) and Kubernetes (K8s)-based deployments are no longer supported for semantic model processing.

• Support for Kubernetes version 1.32: With this release, Kyvos now supports Kubernetes version 1.32 on AWS (EKS) for new deployments.

• Support for JRE version 17 (Spark-based AWS enterprise deployments): Kyvos now supports switching JRE versions between JRE 8 and JRE 17 for AWS Enterprise deployments. For new deployments with Spark or No-Spark on AWS, JRE 17 will be installed by default.

Azure Enhancements

• Load Balancer support for Kubernetes in automated and wizard-based deployment: Kyvos now supports Load Balancer (LB) in Azure for Kubernetes in automated and wizard-based deployments.

• Segmentation in dedicated compute clusters: Kyvos now supports segmentation in dedicated compute clusters. You can create segments in Kyvos Manager and assign semantic models to them for processing. This helps manage resources more efficiently and reduce costs.

GCP Enhancements

• Support for JRE version 17 (Spark-based GCP enterprise deployments): Kyvos now supports switching between JRE 8 and JRE 17 for Spark-based GCP Enterprise deployments. New GCP deployments (with or without Spark) will use JRE 17 by default.

• Support for N4 series machine types: Kyvos now supports N4 series machine types for Kyvos components (BI Server, Query Engine, and Kyvos Web Portal) and the dedicated compute cluster.

• Support for gcloud storage commands: With this release, Kyvos has removed the dependency of ‘aws-cli’ and ‘gsutil’ commands for fresh deployments, as recommended by Google Cloud Platform (GCP).

• Support of No-Spark Shared Query Engine (for Kyvos Free and Kyvos Paid) and Dedicated Compute (only for Kyvos Paid) on GCP Marketplace.

Kyvos Viz. Enhancements

This release comes with the following Kyvos Visualization enhancements:

• Support for dynamic reference lines and bands: Kyvos now supports dynamic reference lines and bands. Label values are automatically updated based on the dynamic caption in calculated measure expressions, making visualizations more interactive and informative.

• Tagging for semantic model elements: You can now tag elements in the semantic model—such as attributes, levels, hierarchies, dimensions, and measures. Tags make it easier to search using the tag filter, organize, and manage these components.

API Enhancements

This release comes with the following new APIs:

1. REST API to fetch Kyvos Manager Audit Logs

2. API enhancements for datasets
   The following new API has been added for datasets.