2.9.2 Release Notes

✨ Release Highlights

• performance improvements

• cluster instances CPU and Memory charts

• admins can see more data in LHO

🖥️ Workloads // All purpose clusters

• 8033 - Cluster instances mem/cpu graphs

  • navigation: select an all purpose cluster → then click on the cluster instance → cpu/memory graphs per instance

  • supports single node and multiple KPIs comparison

  • supports multiple nodes and single KPI comparison

Open

1. click on the name of an All Purpose Cluster

Open

2. Click on the ID of a cluster instance

• Cluster Driver – supports single node and multiple KPIs comparison

Open

3. Use the Cluster Instance View

• Cluster Executor 0 – supports single node and multiple KPIs comparison

Open

• CPU evolution for Cluster Driver and Executor – multiple nodes and single KPI comparison

Open

Security

• 8017 - Change authorization on reports

  • if user is LHO ADMIN or workspace tier is STANDARD then LHO returns ALL data for that user

  • e.g. in other words if I’m an LHO ADMIN, on a Premium Workspace I can see all entities regardless of the DBX authorization rules

  • this change allows the LHO Admin to see resources that were deleted or created by users who are not active anymore (e.g. users deactivated in Databricks) but still have workloads active

Performance Optimizations

• 8265 - Limit OOM dump files size

• 8253 - Improve latency of building the cost estimation report

• 8247 - Reduce default value for SPARK_MAX_NUMBER_OF_CORES

🪩 UX Optimizations

• 8282 - support pagination for cluster instance charts view

Operational Management // Telemetry Analysis

• 8288 - Add JVM stats in the analyzer audit

Miscellaneous

• 8173 - update how the primary and secondary tabs are saved in the URL

  • improve user-friendliness of URL parameters