Rethinking when you test will help you ship faster

Launchable ML learns from code commits & test history meta-data to predict which tests are likely to fail in a test suite—for an incoming set of commits. It then creates a subset of the most-likely-to-fail-tests (think intelligent smoke tests) for these incoming commits. You can then run these subsets earlier in the pipeline—effectively shifting a test suite left.

Thus, developers can find bugs earlier and consequently iterate faster.


Long running tests suites are killing your DX


Nightly runs are usually red and that is a bad DX

Long running test suites implies that they run infrequently. For example, teams usually run a nightly test suite that checks if changes during the day have made the main branch unhealthy.

However, a common outcome is that the nightly test suite is often red, especially true for larger teams.

Thus, the first thing a developer sees when they walk into work is that they broke the nightly build. Their day starts by firefighting instead of focussing on writing new code.

This is terrible DX!


Intelligent subsets keeps nightly test suite green to improve DX

Launchable ML learns from past commit & test suite history to predict which tests are likely to fail in a test suite, for an incoming set of commits.

Teams can then create the Launchable subset based on their nightly test suite and run these subsets throughout the day or on per-commit basis.

Developers do not have to wait till the next morning to see if they turned the nightly test suite run red. They can iterate faster through the day as Launchable smartly flags issues.

Faster iterations implies a great DX.

the curve

Sequential running test suites are killing your DX too.


Sequential test suite runs kill DX and delay releases

Teams have a gauntlet of test suites that a commit runs through. As a commit passes the hurdle of each test suite, the confidence in the release goes up.

However, each of these test suite takes time to run. As a test suite flags issues (multiple days after a commit) the developer has to context switch to fix these issues—for a bad DX.

Issues found too late often start a firefight to assess the impact on the release schedule—often delaying releases.


Intelligent smoke tests— catch issues earlier, improves DX & enables on-time releases

Launchable subsets give teams the ability to create intelligent smoke tests (which effectively moves tests left).

Running the dynamic subset earlier in the pipeline catches issues much earlier thereby improving DX.

This also allows teams the ability to delay the full run of test suites to much later since they have the confidence that intelligent tests are catching issues earlier.


Ready to engage?

Talk to us to see how we can help you go faster

Your cart