Regression Sets

Regression Sets

Regression Sets are reusable collections of test cases and whole sections. Build a set once, then spin up consistent test runs for every build without re-picking cases each time. They are the fastest way to turn "what we test for a release" into a repeatable, one-click run.


Why use Regression Sets?#

  • Repeatable coverage — define a Smoke or Regression suite once and reuse it for every build.
  • Less busywork — no hunting through the library to assemble the same run again and again.
  • Stay in sync — subscribe to a section and new cases added there flow into the set automatically.
  • Consistent reporting — runs built from the same set are directly comparable build over build.

Static vs. dynamic items#

A Regression Set holds two kinds of items, shown on the set's detail page:

Item typeBehavior
Static itemA specific case (or section) you added individually. It is pinned and only changes when you edit the set.
Dynamic caseA case pulled in live from a subscribed section. Add a case to that section and it is automatically included here.

The Current Total on the detail page is the de-duplicated count of static cases plus the current cases from every subscribed section — exactly what a run created right now would contain.

Tip: hover the ⓘ icons on the detail page for an inline reminder of what each number means. Expand a subscribed section row to preview its live cases.


1. Create a Regression Set#

  1. Go to Regression Sets from the project sidebar.
  2. Click New Regression Set.
  3. Give it a name (e.g., Smoke Tests, Release Regression) and an optional description of when to use it.
  4. Add items from the left panel:
    • Cases — pin individual test cases.
    • Sections — subscribe to a whole section so it stays live.
  5. Reorder items to control how they appear in runs, then Save.

2. Create a test run for a build#

This is the core workflow — turn a Regression Set into a run targeting a specific build.

  1. Open the Regression Set and click Create Run.
  2. Name the run (e.g., Sprint 12 — Build 142).
  3. Pick the Release and Build you are testing, plus an environment.
  4. Choose a run mode:
    • Unified — one run containing all resolved cases. Best for a single pass over everything.
    • Split by Set — one run per set. Useful when you create from multiple sets and want them tracked separately.
  5. Click Create Run and start executing.

You can also assemble a run from sets, sections, and individual cases together on the New Run page (Test Runs → New Run) — the Regression Sets and Sections tabs there pull from the same resolution logic.


Useful ways to use Regression Sets#

  • Per-build smoke runs — keep a small Smoke set of P0 cases and create a run against every build for a fast go/no-go signal.
  • Release regression — subscribe whole feature sections so the regression set grows automatically as the library grows.
  • Risk-based suites — build a set of just the cases touching an at-risk area for a focused run before shipping.
  • Recurring cadences — reuse the same set for nightly, per-PR, or pre-release runs so results stay comparable over time.
  • Automation / CI — drive run creation straight from your pipeline via the API (see below).

Automating from CI/CD#

Everything above is available through the API, which is the recommended way to create runs automatically from a build pipeline. See the Regression Sets API reference for endpoints, scopes, and the static vs. section-subscription behavior.