2.16.1 Release Notes

Owned by Ciprian LUCACI
Last updated: Jul 08, 2024
3 min read

 

 

✨ Release Highlights

  • Serverless costs reporting

  • cluster instance incidents & recommendations view

 

Workloads // Workflows

  • 10122 - Include proportional real costs in job reports for jobs using clusters that draw VMs from pools

    • The real costs that we collect for a cluster are of 2 types:

      • normal - from job clusters, that can be composed of ip, disks, VMs and Databricks costs

      • proportional - from jobs that ran on pools, that can be composed only of VMs and databricks costs. 

  • 9164 - set “job name” as default search criteria on workloads overview

image-20240705-165246.png

  • 9716 - display the progress of the subset of values on cluster instances graphs

image-20240705-165427.png
cluster instance view

 

🖥️ Workloads // Job Clusters

  • 9983 - view recommendations per cluster instance id

  • 9960 - cluster instance incidents & recommendations view

 

💠 Workloads // Serverless

  • 9848 - serverless workload costs

  • Jobs running on Serverless

  • details for Serverless Job activity

 

✅ Incidents & Recommendations

  • 10111 - enabled sorting on name, rate, memory and cores on recommendations

 

❤️ Incidents // Performance

  • 10076 - Change the name of incident type into 'Bad Skew'
    Rename "Data Skew" to "Bad Skew" in enum, class names, user messages, queries etc. Impacted incidents can't be updated, and have to be regenerated.

 

🪩 UX Optimizations

  • 10078 - Sort by default descending by Process Memory Max on Clusters Memory Details

    • Allow sorting on memory and cluster dialogs

  • 10100 - India/Mumbai timezone issue on api requests

  • 10113 - Add indentation for lists inside lists on incidents dialog

  • 10098 - added icons and changed design on analysis management navigation

  • 10110 - added sorting to all tables where sorting was not explicitly set to false

 

📈 Operational Management // Consumption (Cost) Analysis

  • 10088 - Refresh overview spend should be called after a consumption loading refresh

 

Miscellaneous

  • 9606 - add unit of measure indicator for time fields for all purpose cluster API

  • 10043 - AWS Dbx token should be allowed as empty

 

🐞 Fixed Bugs

  • 10083 - No estimated costs for job runs

  • 10033 - in Cluster Instances view the app id is sometimes wrongly marked as "unknown"

  • 10095 - Trends API not using estimated costs to compute Cost per data/ Cost per duration rates

  • 10048 - Incidents screen - kebab button is overlapped to incident time when screen width under 1200 pixels

  • 10045 - Not able to open app cluster instance in another tab using cmd + click