Change Log
New Features
Import Export - Workspaces/Jobs/Notebooks
Introducing the ability to import and export Workspaces, Jobs, and Notebooks, enabling seamless sharing and migration of assets across environments.
Multiple Filter Support - Multiple Status to Filter
Support for multiple status filters has been added to enable more granular data segmentation. Users can now apply multiple status filters simultaneously to get a refined view of their data inside workspaces dashboard.
Thrift Server
The Thrift job type enables users to start a Thrift SQL session to execute SQL queries via a driver. This driver can be integrated with tools like DBeaver, allowing for seamless SQL query execution.
Yeedu Turbo for Spark-3.4.3
Users can now enable Turbo Mode with Spark 3.4.3, designed to optimize performance for compute-intensive workloads. Turbo Mode enhances efficiency by accelerating job execution while minimizing resource consumption, making it ideal for handling demanding data processing tasks.
Version Upgrade Scripts
Starting from the v2.7.0 version of Yeedu, users can upgrade to the latest stable versions without resetting the database. Upgrade scripts can be executed to ensure a smooth transition while retaining all existing data.
Support for New Spark Versions
Yeedu now fully supports Spark versions 3.5.1 and 3.4.3, providing compatibility with the latest features and improvements in these releases.
Minor Enhancements
OSM Synchronization Enhancement to Reduce Memory Load
Optimization of the OSM synchronization process, resulting in reduced memory usage. This enhancement improves system stability, especially for resource-intensive operations.
2. Restart VM Initialization if Bootstrap Fails
In cases of machine terminates or restarts from cloud the machine will properly bootstrap and job executions resume as per design.
4. Workflow Job Directory Management
Jobs are now executed within directories created based on workflow_job_instance_id and are automatically cleaned up according to a defined cleanup timeout. This structure prevents clutter, optimizing storage usage and organization.
5. Cluster Autoscale Enhancement
Cluster autoscaling has been improved to respond even when a single job is queued for execution. This change ensures that resources are always available to meet demand, reducing wait times for job execution.
6. UX Enhancements
Various user interface improvements have been implemented across the platform, creating a smoother and more intuitive experience for users.
7. Autoscaling Based on history_cluster_instance_id
Auto start, resize, and downscale operations now utilize history_cluster_instance_id instead of compute_engine, improving resource tracking and scaling precision.