Feature Requests

Feature request : The ability to rerun builds and workflows multiple times while they are still running. This allows me to rule out flaky tests. Sometimes, I need to run the same workflow multiple times at the same time to rule out flaky test. This used to be possible by selecting "rerun" on the currently running build. That option is now greyed out while the build is running. This was possible in the old UI and it was very useful. CCI-I-1614

Flakey tests are currently tagged with "Flakey". We want to be able to manually mark these test as "fixed" or to remove the "Flakey" tag so that we know that the test is expected to now pass. If it does not, the "Flakey" test tag will automatically be re-added to the test.

Currently, test splitting by timing uses the following fallback logic: Latest job's test results from the same branch and if no data exists, falls back to latest job's test results from ANY branch in the project. Requested Feature: Add configuration option to specify which branch's timing data should be used for test splitting, such as: - circleci tests split --split-by=timings --timing-branch=main Or a configuration at the job level: jobs: test: restore_test_results: timing_source: main # Only use timing data from main branch Why This Matters: Bugs or performance issues in PR builds shouldn't corrupt the timing baseline for all subsequent builds Partial test runs in branches produce incomplete timing data that leads to poor test distribution Teams need predictable, reliable test splitting based on a stable timing baseline

Knowing if a test is flaking on main or dev branches can be helpful when attempting to diagnose and/or report on flaky tests. To know which branch a flaky test is on, two API calls need to be made currently. It would be helpful if the branch name was associated with the test flake instance within the flaky test endpoint.

I'm interested to know how often people in our organization are having to rerun jobs when a flaky test fails on their branch. Who performed the rerun, when was it done, which repo, which branch, which job, which test failed, etc.

Currently this is only available via the GUI. > https://circleci.com/docs/rerun-failed-tests/ This feature request is to introduce this functionality to users via a CircleCI API endpoint

Currently I can only submit screenshots only as artifacts (with store_artifacts command) and see them in artifacts section like in attachments (see scan-test-results contents). Screenshots usually are stored in "<derived data folder>/Logs/Test/Attachments" folder, and have UUIDs in their names instead of understandable name which could be used to match test and test step with the screenshot. Information about this matching containing in "<derived data folder>/Logs/Test/*_TestSummaries.plist" and is easy to parse. It would be awesome if there will be some separate command like store_uitest_results which will receive test results folder as a parameter (or even do it automatically), parse plist file and create a Web page with all screenshots split by tests and ordered by their timestamp. CCI-I-406

When viewing tests failed (using RSpec), it has a section listing "X tests failed out of Y" and gives the name of each test with a "copy to clipboard" icon. However the name is not useful when wanting to rerun that test locally. Could a "copy to clipboard" action be added to the paths listed below where the details of each test are given, including the file path?

It costs time and mental energy to enter the test job first, scroll up and then click the "Rerun" dropdown.

In the UI to see flaky test data a workflow and a branch needs to be selected on the Tests tab in insights. This makes it difficult to attribute flakiness across the project and track that flakiness over time. It would be helpful if there was flaky test information about a project at a glance and if that could be drilled down into without needed to select specific branches first.