For Developers‎ > ‎

Cluster Telemetry

Run on Cluster Telemetry! (Google only, sorry)

Cluster Telemetry 101

Telemetry is Chrome's performance testing framework, using it you can perform arbitrary actions on a set of web pages and report metrics about it.

Cluster Telemetry allows you to run telemetry's benchmarks using multiple repository patches through Alexa's top 10k and top 100k web pages.
Developers can use the framework to:
  • Measure the performance of multi-repo patches against the top subset of the internet on Android devices, bare-metal linux machines, and GCE instances using the perf page.
  • Gather metrics for analysis or for reports against 100k web pages on CT's VM farm (linux and windows) using the analysis page.
  • Compute metrics on an analysis run using the metrics analysis page.

Should I use CT?

If you would like to know how your patch impacts Chrome's performance or would like to do analysis against a large repository of real world web pages then you should try out Cluster Telemetry.

CT has been used to gather analysis data for the following projects:
  • Web vitals performance
  • Ad tagging accuracy
  • UserCounter analysis
  • Real-world leak detector
  • Gather data for a language detector
  • TTI variability study
  • Analysis of 5M SVGs
  • Stat rare usages
CT has been used to gather perf data for the following projects:
  • Slimming paint
  • Clustering top web sites for Chrome graphics
  • Performance data for layer squashing and compositing overlap map
  • SkPaint in Graphics Context
  • Culling
  • New paint dictionary

Which telemetry benchmarks does CT support?

CT supports all benchmarks listed in tools/perf/contrib/cluster_telemetry. Many of these benchmarks exist outside of CT, but these are the corresponding CT compatible versions.
CT also allows you to run against unlanded / modified benchmarks using the multi-repo patches supported in the UI.

Can try out your benchmark's CT version locally with:
python tools/perf/run_benchmark ${YOUR_BENCHMARK} --also-run-disabled-tests --user-agent=desktop --urls-list= --archive-data-file=/tmp/something --output-dir=/tmp/output/ --browser-executable=/usr/bin/google-chrome --browser=exact --pageset-repeat=1 --output-format=csv --use-live-sites

How accurate are CT's results?

For an empty patch repaint run on Desktop, these are the results:


The overall results from Cluster Telemetry runs are accurate within a percentage point.

The per webpage results (visible when you click on a field) do have some variance, but this has been greatly improved due to efforts detailed here.

Framework Code and Documentation

Cluster Telemetry is primarily written in Go with a few Python scripts. The frontend is written using Polymer. The framework lives in main/ct.
Here is more detailed documentation for the different pages:

Contact Us

If you have questions, please email rmistry@ or