Classic Docs
Docs Navigation

Interpreting Test file performance data

After you have enabled Test file performance Insights on a project and you've ran a successful build, you will be able to access a report on your slowest test files. You can see the report by going to the project page and clicking the Insights link, and then opening the Test file performance tab.

You can choose the framework for which you want to see the report from the Insights page. For example, if you want to see which RSpec files are the slowest in your test suite, if you follow the RSpec link, you will see a table similar to the one in the screenshot below.

RSpec table

# Report data

The table contains all test files which take more than a minute to run, and if there are no files that fit that description, then the slowest five are shown by default.

It also includes the maximum and average running time of the listed files. You can also see how often a test file passed.

Each table row can be expanded to show more info on a file's build history. Bar height indicates running time, and the build outcome is color-coded. The data in the report is based on the last 50 builds.


# How to use Test file performance data to enhance your tests

If you notice some test files with a lower build pass rate and you're not sure why that keeps happening, it's possible that your test suite contains a flaky test. You can check out this article to find out how you can eliminate them.

There can also be some test files which are significantly slower than the rest of the suite. This can happen with tests that make excessive database calls, or are managing resources in a similar inefficient manner.

This can become an issue if you have a test file that takes 5 minutes to complete, which means that your whole build cannot complete in less than 5 minutes, plus the time it takes to perform any setup and teardown operations that may exist for your project.

High values in Standard deviation column usually indicate there is potential for optimization. Semaphore Community has some articles which can help you learn more about writing good specs:

# Further reading

Semaphore Docs are open source — Edit on GitHub


Occasional lightweight product and blog updates. Unsubscribe at any time.

2009-2021 © Rendered Text. All rights reserved. Terms of Service, Privacy policy, Security.