Khan / jenkins-perf-visualizer

Visualize how your Jenkins builds are spending their time

jenkins-perf-visualizer

Visualize how your Jenkins builds are spending their time

Preface

jenkins-perf-visualizer was developed for use by Khan Academy and is being released publicly in the hopes it will be useful to others. It is not being actively maintained, and Khan Academy will resolve any filed issues on a best-effort basis only. (Pull requests are more likely to be considered.)

Khan Academy is a non-profit organization with a mission to provide a free, world-class education to anyone, anywhere. You can help us in that mission by donating or looking at career opportunities.

Overview

This project provides various tools for analyzing and visualizing what your Jenkins builds are doing over time.

Jenkins has some built-in visualization support, and there are also visualization plug-ins that work inside Jenkins.

This tools differs from those in the following respects:

1. It provides datapoints not only on a per-build basis, like most plug-ins, or on a per-stage basis, like the built-in visualizer, but also per-parallel-step. This makes it easy to see, when running various tasks in parallel, which is the critical path.

2. It shows time-taken visually, as a stacked bar graph. Other tools show time-takes as a number, making it hard to see at a glance where the time is being spent.

3. It can graph a single "task" -- a collection of different build jobs that are part of a workflow -- on a single graph. Other tools can only visualize a single build at a time, or, at best, historical builds for a single job.

4. It is a standalone tool, not part of Jenkins. Among other advantages, this makes it easy to examine historical data after Jenkins has deleted the information about a build.

5. The raw data used to generate the graph is stored in an easy-to-parse format, making it easy to run analyses over the data, use the data for alerts, etc. For instance, in the example graph below, you could easily compute for each of the deploy-to-x nodes how often it was the one that took the longest to run.

6. It distinguishes between time a build spends running, vs waiting for user input, sleeping, or waiting to start (because no executor is available). This gives valuable insight how to solve performance problems.

Here is an example graph:

Getting Started

TODO

Configuring jenkins-perf-visualizer

TODO

Implementation Details

jenkins-perf-visualizer uses the Jenkins API to get some information about jobs and builds, but mostly depends on the Jenkins "Pipeline Steps" page.