Runs and Logs
A run is a single execution instance of an E2E test or browser agent session. It captures the full lifecycle: trigger, browser actions, screenshots, logs, and final outcome.
Triggers: Manual execution, CI/CD webhooks, scheduled jobs.
Global Runs List
/e2e-runs shows all runs across every test in the project.
Stats bar: Total Runs (with success rate %), Running, Pending, Passed, Failed.
Filter tabs: All / Running / Pending / Passed / Failed
Table columns: Test name, type, started by, created, status.
Advanced filters: Run type (Manual / Scheduled / Triggered), date range, started by. Expand any row to see inline details without leaving the list.
Status Reference
| Status | Meaning |
|---|---|
| Running | Currently executing |
| Pending | Queued |
| Passed | All assertions passed |
| Failed | Assertions failed or error during execution |
| Cancelled | Manually stopped |
| Timeout | Exceeded time limit |
Run Detail Page
Open a run from the global list or from a test's Test Runs tab.
Tabs: Overview, Conversations.
Overview shows: run ID, test name, status and outcome, started by, created timestamp, duration, step count, target URL, and run type. Includes the test recording (if available).
Conversations shows the full agent conversation log from the run.
For step-by-step playback, use the Run Viewer.
Log Viewer
The Log Viewer is used for pipeline and repository runs that produce Docker build output.
Two tabs:
- Build Logs — Dockerfile generation and Docker build output
- Run Logs — Container execution output
Features:
- Terminal styling: black background, green text, red for stderr
- Line numbers on the left
- Copy all logs button
- Infinite scroll — loads 100 lines at a time; scroll up to load older lines
- "── Beginning of logs ──" marker when you reach the start
The run header shows: commit SHA, branch, status badge, created timestamp, and a refresh button.
Failure Insights
For failed pipeline runs, a Failure Insights card appears automatically containing:
- Failure mode — categorized cause (e.g., "Navigation Failed", "Assertion Failed")
- Confidence — how certain the AI is (e.g., 94%)
- Description — plain-language explanation of what went wrong
- Suggested fixes — actionable steps to resolve the issue
- Failure Timeline — chronological list of diagnoses with supporting evidence