spark-sight: Spark performance at a glance

spark-sight is a less detailed, more intuitive representation of what is going on inside your Spark application in terms of performance.

Read more on Medium:

1. Part 1: Meet "spark-sight": Spark Performance at a Glance
2. Part 2: “spark-sight” Shows Spill: Skewed Data and Executor Memory

What is it?

spark-sight is a less detailed, more intuitive representation of what is going on inside your Spark application in terms of performance:

• CPU time spent doing the “actual work”
• CPU time spent doing shuffle reading and writing
• CPU time spent doing serialization and deserialization
• Spill intensity per executor (v0.1.8 or later)
• (coming) Memory usage per executor

spark-sight is not meant to replace the Spark UI altogether, rather it provides a bird’s-eye view of the stages allowing you to identify at a glance which portions of the execution may need improvement.

Main features

The Plotly figure consists of charts with synced x-axis.

Top chart: efficiency

The top chart shows efficiency in terms of CPU cores available for tasks

Middle chart: spill

The middle chart shows spill information

Bottom chart: stage timeline

The bottom chart shows stage timeline

Where to get it

pip install spark-sight

Dependencies

• Pandas: Powerful Python data analysis toolkit for support with Spark event log ingestion and processing
• Plotly: The interactive graphing library for Python for the awesome interactive UI

Usage

spark-sight --help

                      _             _       _     _
 ___ _ __   __ _ _ __| | __     ___(_) __ _| |__ | |_
/ __| '_ \ / _` | '__| |/ /____/ __| |/ _` | '_ \| __|
\__ \ |_) | (_| | |  |   <_____\__ \ | (_| | | | | |_
|___/ .__/ \__,_|_|  |_|\_\    |___/_|\__, |_| |_|\__|
    |_|                               |___/

usage: spark-sight [-h] [--path path] [--cpus cpus] [--deploy_mode [deploy_mode]]

Spark performance at a glance.

optional arguments:
  -h, --help            show this help message and exit
  --path path           Local path to the Spark event log
  --cpus cpus           Total CPU cores of the cluster
  --deploy_mode [deploy_mode]
                        Deploy mode the Spark application was submitted with. Defaults to cluster deploy mode

Unix

spark-sight \
    --path "/path/to/spark-application-12345" \
    --cpus 32 \
    --deploy_mode "cluster_mode"

A new browser tab will be opened.

Windows PowerShell

spark-sight `
    --path "C:\path\to\spark-application-12345" `
    --cpus 32 `
    --deploy_mode "cluster_mode"

A new browser tab will be opened.

Read more on Medium

spark-sight Medium series, part 1

Meet "spark-sight": Spark Performance at a Glance

spark-sight Medium series, part 2

“spark-sight” Shows Spill: Skewed Data and Executor Memory