Dr. Estera Kot

ML based Autotune for Apache Spark Jobs in MS Fabric performance optimization for recurrent jobs

Episode 65 of Fabric Espresso, hosted by Dr. Estera Kot.

Video thumbnail for ML based Autotune for Apache Spark Jobs in MS Fabric performance optimization for recurrent jobs. Opens the video on YouTube.

Length 28:11. Published Aug 13, 2024 on the Azure Synapse Analytics YouTube channel. Topic: Spark and Data Engineering.

Watch on YouTube

About this episode

Text from the YouTube description.

In this episode of Fabric Espresso, we delve into Autotune. Estera Kot, along with Karuna Sagar Krishna and Yiwen Zhu unpack the technology, use cases and demo of Autotune – predictive optimization for Apache Spark Jobs - which automatically adjusts Apache Spark configuration to speed up workload execution and to optimize overall performance. Autotune configures three Apache Spark settings for each of your queries separately based on historical runs.

Read more: https://learn.microsoft.com/en-us/fabric/data-engineering/autotune

🎙 Meet the Speakers:

👤 Guest Expert: Karuna Sagar Krishna | Principal Software Engineering Manager at Microsoft

LinkedIn:https://www.linkedin.com/in/karuna-krishna-842211126/

👤 Guest Expert: Yiwen Zhu | Principal Applied Data Scientist, Researcher at Microsoft

LinkedIn: https://www.linkedin.com/in/yiwen-zhu-25139933/

👤 Host: Estera Kot, PhD, Principal Product Manager at Microsoft.

LinkedIn: https://www.linkedin.com/in/esterakot/

Twitter: https://twitter.com/estera_kot

👍 Liked this video? Don't forget to hit the 'Like' button and share it with your peers!

Chapters

  1. 0:00 Introduction
  2. 1:55 Spark settings
  3. 7:30 Autotune performance
  4. 13:20 Technical details
  5. 20:07 AI based intuition
  6. 20:23 Autotune approach
  7. 21:25 Finetuning process
  8. 22:44 Centroid update
  9. 25:06 Enable Autotune

All Fabric Espresso episodes