Running Tests

Once you have generated test cases from your document, it is time to run them. This guide covers selecting browsers, starting runs, monitoring progress, and what happens when tests need to self-heal.

​

Selecting Browsers

Marketrix QA supports three browser engines: By default, tests run on Chromium. You can configure your document’s browser settings to run across multiple browsers:

• Single browser — Pick one and run. Fastest option.
• Multiple browsers — Select any combination of Chromium, Firefox, and WebKit.
• Parallel execution — Run all selected browsers at the same time instead of sequentially.

​

Configuring Browser Settings

On your document’s detail page, look for the browser configuration section. You can set:

• Which browsers to include
• Whether to run in parallel
• Whether to stop all browsers on the first failure (fail fast mode)
• Timeout per browser (between 60 seconds and 1 hour)

​

Starting a Test Run

1. Navigate to your document in Quality Checking
2. Click Run Tests or Run Cross-Browser
3. Confirm your browser selection
4. The run starts immediately

Each run creates a record you can revisit later. Every test case in the document is executed as part of the run.

​

Monitoring Progress

While tests are running, the dashboard shows real-time progress:

• Overall run status — Pending, Running, Completed, or Failed
• Individual test status — Each test case shows its own progress
• Step-by-step log — See which step each test is on, with timestamps
• Live updates — The page refreshes automatically as tests progress

​

Understanding Run Statuses

​

Understanding Test Statuses

Individual test cases within a run have their own statuses:

​

Stopping a Running Test

If you need to stop a test run in progress, you can cancel it from the run’s detail page. Any completed tests within the run will keep their results. Tests that were still pending or running will be marked accordingly.

​

Self-Healing: Automatic Test Repair

This is one of the most valuable features of Marketrix QA. When a test fails because your UI has changed — a button moved, a selector changed, a form field was renamed — the system can automatically attempt to fix the test.

​

How Self-Healing Works

1. Failure detected — A test fails due to a missing element, changed selector, or broken assertion
2. Analysis — The system identifies the type of failure (locator, assertion, timeout, flow change, or environment issue)
3. Repair attempt — A new approach is generated with a confidence score
4. Validation — The repair is tested to verify it actually works
5. Approval workflow — You review and approve or reject the fix

​

Failure Types the System Can Heal

​

Reviewing Healing Attempts

When a self-healing attempt is made, you will see it in the test’s healing history. Each attempt shows:

• What failed and why
• The proposed repair strategy
• A confidence score (0 to 1) — higher is better
• The validation status (pending, validated, failed, or rejected)

You can approve a healing attempt to apply the fix permanently, or reject it if the repair does not look right. This keeps you in control — the system suggests fixes, but you decide what sticks.

​

Healing Configuration

You can tune healing behavior per test:

• Enable/disable healing — Turn self-healing on or off
• Auto-apply — Automatically apply high-confidence repairs without manual approval
• Confidence threshold — Set the minimum confidence score for repairs (default: 0.85)
• Max attempts — Limit how many healing attempts per failure (default: 2, max: 5)

​

Troubleshooting

Tests are stuck on “Pending”:

• Check that your connection is active and the target URL is accessible.
• Verify your plan supports the number of concurrent runs you are attempting.

All tests fail immediately:

• Your connection credentials may have expired. Update them on the Connections page.
• The target application may be down. Try accessing the URL manually.

Tests pass on Chromium but fail on Firefox/WebKit:

• This usually indicates a real cross-browser compatibility issue in your application.
• Check the screenshots to see what looks different.
• These are legitimate bugs worth investigating.

Self-healing keeps failing:

• The UI change may be too large for automatic repair. Manually update the test steps.
• Try re-uploading an updated document to regenerate the affected tests.

​

Next Steps