2.29 Release Notes

2.29 Release Notes

 

Sep 4, 2025

Lakehouse Optimizer 2.29.2

 

Audio Overview of Release

 

Cost Over Time

  • support weekly and daily cost breakdowns

    • support cost breakdown aggregated per week or per day

    • enhanced navigation from Forecasting page when investigating daily costs spikes

image-20250904-131820.png
by week
image-20250904-131843.png
by day

 

Cost Analysis

  • report AI Gateway cost in Cost Over Time breakdown Feature

image-20250904-130837.png

 

Forecasting page

  • Weekly spend KPI chart

image-20250904-125823.png
  • Forecasting: Cost & Data Usage chart: Avg. KPI value for the selected time period

    • each selected KPI has an averate for the selected period

image-20250904-131036.png
image-20250904-131117.png
image-20250904-131136.png

 

Vendor Consolidation

  • improved documentation when data is not available, for N/A scenarios

image-20250904-125709.png

 

Provisioning Workloads

  • map legacy_single_user_standard data mode to 'single'

image-20250904-125308.png

 

Data Analysis

  • enhance orphaned task metrics resolving for job cluster analysis

  • improve performance on job runs trends page

  • add Data In, Data Out, Shuffle In, Shuffle Out in Edit table columns > Cluster instances page

image-20250904-131609.png

 

Navigation

  • link RunJob Tasks to the actual Job definition

    • workflows that have tasks as RunJob include now a link from the task to the actual RunJob parent

image-20250904-131331.png
click on the run ID to open a new tab with the parent job
image-20250904-131351.png

 

Documentation

  • tooltip for "Auto Stop" column in SQL Warehouses table with link to Databricks article

    2025-09-05_15h41_34-20250905-224134.png

 

Miscellaneous

  • updated design for job run statuses

  • list workspaces and subscriptions from DB instead of using cloud provider API

 

 

Aug 27, 2025

🚀 Lakehouse Optimizer 2.29.1

Dependent Vendor summary per job run and task

Updated design for exposing vendor informaion at task level and per job run.

  • The total data transferred and total wait time is the aggregation of unique Spark Job Groups metadata.

  • Data Transferred and Wait Time are extracted from Spark Job Group metadata. If multiple vendors are used simultaneously in the same Spark Job Group, then it is not possible to distinguish the direction and duration of data flow attributed to each unique vendor. In this scenario, the entire Spark Job Group’s values for data and time attributed to each vendor.

image-20250827-121917.png
dependent vendor summary
image-20250827-121729.png
dependent vendor breakdown per job run
image-20250827-121807.png
dependent vendor breakdown per task

 

Support global date range across Workloads pages

  • once date is set on one page in workloads, the date will be passed also to all pages when navigating, unless there is another contraint, such as the exact date of a run

 

Enhanced documentation on Forecasting: Cost & Data Usage widget

  • add short description what is Data usage and Data Usage Cost

image-20250827-122622.png
image-20250827-122646.png

 

Dependent Vendor information for Notebooks

image-20250827-122919.png

 

Dependent Vendor information for Pipelines

image-20250827-122944.png

 

Job Runs Trendline KPIs documentation

  • each KPI includes now a description

image-20250827-123158.png
image-20250827-123122.png

 

Filter Migration Opportunities by Vendor

image-20250827-123343.png

 

Forecasting : update CSV export feature to include all KPIs

 

 

Aug 20, 2025

🚀 Lakehouse Optimizer 2.29 – What's New

🔍 Cost and Forecasting Enhancements

  • support Enhanced Security and Compliance cost category to Databricks Platform

    • Cost Over Time > Databricks Platform > Enhanced Security and Compliance

image-20250820-152716.png
image-20250820-152803.png
  • Forecasting Weekly Fixes

    Weekly cost KPIs and reports now follow ISO 8601 week numbering (Monday–Sunday), respecting user timezones. UI and CSV exports are now aligned.

image-20250820-162249.png
  • Forecasting KPIs Improvements

    Refined KPI groupings on the Forecasting page, improving clarity and structure for monthly views.

image-20250820-162124.png
  • Cost Over Time CSV Fixes

    CSV exports now include all relevant cost breakdowns. Improvements ensure consistency across breakdown levels and export formats.

  • KPI: Run Estimated Cost Visibility

    The “Run Estimated” cost column is now always visible in Job Run views, helping users monitor cost projections directly.

  • KPI Calculations Fixes

    Addressed several inconsistencies in forecasting KPIs and cost metrics (e.g., Cost per TB, Data Usage), including edge cases and null handling.

  • Improved Forecasting Tooltips and Layouts

    Adjusted hover interactions and bar chart tooltips for clarity and responsiveness. Enhanced UI for forecasting side panels.

  • Breakdown for Other DBU/Cloud costs in Cost Over Time report

image-20250820-155439.png
  • Display estimate badge for job cluster costs

    A new badge visually flags when job cluster cost values are estimated, not actual.

  • Enable correlation for cost across pipelines with shared tasks

    Enhanced cost attribution across pipelines with shared components for more accurate reporting.

  • Fix rounding issue in forecast chart

    Spend Forecasts now display rounded values more precisely.

  • in Workloads: Jobs > Trend - Tasks add support for "Duration - Setup Only" and "Duration - Cleanup Only" KPIs

    • enables you to compare for each task how much it takes for the cluster instance to get setup or shutdown

image-20250820-153807.png

 

 

🧠 Recommendations & Incidents

  • Persist job run recommendations

    Recommendation history is now saved per job run, allowing users to track past optimization suggestions and results.

  • Fix search in historical recommendations

    Historical recommendation queries now use the incident date instead of run start time, improving accuracy when searching.

  • Update DBX query for jobs with APCs incidents

    Backend queries were refined for improved detection of incidents on All-Purpose Clusters (APCs).

  • Expose feature flag status for autopilot

    Autopilot recommendation engine status is now visible to administrators for better monitoring and transparency.

  • Support rule override per workspace (Scope section)

    Users can now override SKU Cost incident rules at the subscription or workspace level, giving greater control over incident thresholds and alerts.

  • Scope documentation for policy configuration panel

image-20250820-155051.png
  • Recommendation Enhancements

    The Disk Spillage recommendation now includes the spilled data size directly in its title for easier prioritization.

image-20250820-154849.png
  • Autoscaling Timeline Link for Recommendations

    Now available for Disk Spillage and Bad Skew insights, offering direct traceability to resource scaling events.

image-20250820-154507.png
  • support Incident to compare daily cost of SQL Warehouse with AVG daily cost of Last Week, Last Month

image-20250820-161449.png

 

📈 UI and UX Improvements

  • Unified Widget Proportions

    Harmonized widget proportions on Forecasting and Vendor Consolidation pages for a more consistent and readable UI.

image-20250820-161254.png
image-20250820-161311.png
  • Session Expired Window

    Improved UX for expired sessions with clearer messaging and visual indicators.

  • Cost Type Badges on Trend Views

    Badges indicating Real/Estimated cost types were added or corrected across several trend pages for consistency.

  • Fix tab name when navigating from recommendations to workloads

    Preserves the correct tab context during navigation for consistent user flow.

  • enhanced workloads navigation

    • a UI spacing fix improves visual alignment

    • add "Next" and "Back" words for page navigation with enabled and disabled states

  • Fix trendlines error for empty workloads

    Prevented UI failures when loading trendlines for very short or idle workloads.

  • Improve dropdown display speed and styling

    Faster dropdown responsiveness and improved layout consistency.

  • Fix double cursor pointer in search boxes

    Resolved overlapping cursors in input fields for a cleaner experience.

 

 

🧮 Data Usage & Vendor Analytics

  • New Cost Categories

    Added support for tracking Databricks Platform / Serverless GPU Compute and Enhanced Security & Compliance costs.

  • Data Usage Insights

    Rolled out data usage statistics on cluster instance views and improved the “Cost of Data” KPI formula (now cost / data, using correct units).

image-20250820-162153.png
  • AI & Model Training Cost Tracking

    Foundation Model training costs and AI “other” expenses are now categorized and displayed correctly across reports.

  • Vendor Info at Task Level

    Granular vendor metadata (including source database or cloud provider) now appears at the task level within workflows.

image-20250820-162027.png
image-20250820-162047.png

 

  • Vendors filtering on job runs

image-20250820-155228.png
image-20250820-155255.png
image-20250820-155318.png

 

 

⚙️ Engine & Infrastructure Updates

  • Autoscaling & Serverless Adjustments

    Enhanced logic for identifying serverless runs and computing proportional costs accurately across retries and multi-attempt jobs.

  • Job Config Timeline

    Repaired inconsistencies in fetching job config snapshots for time-based analyses.

  • Query Metadata Expansion

    Job query references now include notebook path, ADF flag, and start time, allowing for deeper insight into job context.

  • APC Cost Calculation Enhancements

    Optimized how execution duration is factored into estimated APC (All-Purpose Compute) costs.

  • Fix issue when cluster_id is missing in runtime analysis

    Resolved failures in job run analysis when certain cluster identifiers were unavailable.

  • Correct job run metadata when cluster name is missing

    Ensures job runs without named clusters still retain accurate metadata.

  • Add DBX workspace ID to job runs table

    Workspace IDs now appear in job run tables, supporting traceability across environments.

 

 

🧪 Stability and Bug Fixes

  • Add success criteria logs to help with debugging

    Success criteria checks are now logged to simplify support investigations.

  • Efficiency Incident Validations

    Fixed validation issues on incident rules related to excessive duration, idle time, and auto-shutdown parameters.

  • Visibility enhancements

    • Fixed incorrect default sort in Cost Controls.

    • Corrected layout spacing and alignment issues across Workflow Runs, Jobs Trend, and other views.

    • Adjusted tooltip spacing and icon visibility on multiple pages.

  • Technical Debt Reduction

    • Deprecated old reports API code.

    • Simplified cost pipeline computations.

    • Improved proportional cost calculations across shared clusters.