Smart Test Selection

Instead of running every test case every time, use smart selection strategies to run only the tests that matter. This cuts 30-minute full runs down to minutes — ideal for CI pipelines and quick feedback loops.

Available Strategies

Strategy	What It Runs	Best For
all	Every test case in the project	Nightly / full regression runs
recent-failures	Cases that failed in the last 7 days	Quick re-validation after fixes
high-priority	Cases marked as high or critical priority	Critical path testing
smoke	One representative case per page URL	Fast sanity check across all pages
changed-routes	Cases matching a URL pattern you provide	Testing specific areas after deployment
regression-risk	Cases with high priority + recent failure history (excludes known flaky tests)	Targeted regression prevention

Using Smart Selection in the UI

Using Smart Selection via API

List Available Strategies

GET /api/v1/projects/:projectId/selection-strategies

Response:
{
  "strategies": [
    { "key": "all", "label": "All Tests", "description": "Run every test case", "estimatedCount": 4858 },
    { "key": "recent-failures", "label": "Recent Failures", "description": "Cases that failed in last 7 days", "estimatedCount": 1683 },
    { "key": "high-priority", "label": "High Priority", "description": "Critical and high priority cases", "estimatedCount": 0 },
    { "key": "smoke", "label": "Smoke Tests", "description": "One case per page URL", "estimatedCount": 438 },
    { "key": "changed-routes", "label": "Changed Routes", "description": "Cases matching a URL pattern", "estimatedCount": null },
    { "key": "regression-risk", "label": "Regression Risk", "description": "High-risk cases based on priority and failure history", "estimatedCount": 0 }
  ]
}

Create a Run with a Strategy

POST /api/v1/test-runs
{
  "projectId": "your-project-id",
  "runType": "project",
  "selectionStrategy": "smoke"
}

The backend resolves the strategy to specific case IDs before executing. The run header will show a smoke badge to indicate which strategy was used.

Changed Routes Strategy

The changed-routes strategy requires a urlPattern parameter:

POST /api/v1/test-runs
{
  "projectId": "your-project-id",
  "runType": "project",
  "selectionStrategy": "changed-routes",
  "selectionParams": {
    "urlPattern": "/checkout"
  }
}

This runs all test cases whose page URL contains /checkout.

CI/CD Integration

Smart selection works with the CI trigger endpoint. Use different strategies for different pipeline stages:

GitHub Actions Example

# PR checks - fast smoke tests
- name: Run Smoke Tests
  run: |
    curl -X POST https://aegisrunner.com/api/v1/ci/trigger \
      -H "Authorization: Bearer ${{ secrets.AEGIS_API_KEY }}" \
      -H "Content-Type: application/json" \
      -d '{"projectId": "YOUR_PROJECT_ID", "selectionStrategy": "smoke"}'

# Nightly - full regression
- name: Run Full Tests
  run: |
    curl -X POST https://aegisrunner.com/api/v1/ci/trigger \
      -H "Authorization: Bearer ${{ secrets.AEGIS_API_KEY }}" \
      -H "Content-Type: application/json" \
      -d '{"projectId": "YOUR_PROJECT_ID", "selectionStrategy": "all"}'

GitLab CI Example

smoke_tests:
  stage: test
  script:
    - |
      curl -X POST https://aegisrunner.com/api/v1/ci/trigger \
        -H "Authorization: Bearer $AEGIS_API_KEY" \
        -H "Content-Type: application/json" \
        -d '{"projectId": "YOUR_PROJECT_ID", "selectionStrategy": "smoke"}'
  rules:
    - if: $CI_PIPELINE_SOURCE == "merge_request_event"

full_regression:
  stage: test
  script:
    - |
      curl -X POST https://aegisrunner.com/api/v1/ci/trigger \
        -H "Authorization: Bearer $AEGIS_API_KEY" \
        -H "Content-Type: application/json" \
        -d '{"projectId": "YOUR_PROJECT_ID", "selectionStrategy": "all"}'
  rules:
    - if: $CI_COMMIT_BRANCH == "main"
      when: always

Strategy Badge on Run Results

When viewing a completed run that used a selection strategy, a purple badge appears in the run header showing which strategy was used (e.g., smoke or recent-failures). This helps you distinguish targeted runs from full regression runs.

Combining with Failure Triage

Smart selection and failure triage work together. After a targeted run completes with failures, triage automatically analyzes just those failures. This gives you focused insights for the specific area you tested.