Record browser tests — video, steps, HAR & console

When your test drives a real browser, a pass/fail number isn’t enough — you need to see what the browser did. For every browser load test, MaxoPerf records four things per browser instance and ties them together on the run-detail Video tab:

• Video — a clean recording of the browser (no “controlled by automated test software” infobar).
• Steps — the ordered actions the test took (navigate, find element, click, type, …), each timestamped and synced to the video.
• HAR — the full network waterfall (requests, responses, status, size, timing).
• Console — the browser console output (logs, warnings, errors).

Tip

This is automatic for browser tests. You do not edit your test script, add logging, or configure a proxy — MaxoPerf instruments the browser on the runner for you.

What counts as a browser test

Any test whose executor drives a real browser through Selenium WebDriver — including a Taurus YAML scenario with requests:/actions:, or your own Python Selenium script (webdriver.Chrome(...)). Capture is on by default; non-browser (API/protocol) tests are unaffected and incur zero overhead.

Today: Chrome/Chromium gets full fidelity (video, steps, HAR, console). Firefox and the other browser engines are on the roadmap — see Coverage.

View it on a run

1. Open the run, then the Video tab.

2. If the run used more than one browser instance (multiple virtual users, or a test that re-launches the browser per iteration), pick the one you want from the browser selector. Each instance has its own steps, HAR, and console.

3. Watch the step timeline. The step that was running at the current video time is highlighted. Click any step to jump the video to that exact moment — e.g. click “type into #password” to see the form being filled.

   

   Run detail Video tab — the step timeline stays in sync with the recording; click a step to jump the video to that moment.

4. Open the HAR view for the network waterfall — filter by type/status, inspect request/response headers, timing, and sizes, or download the raw HAR.

5. Open the Console view for the browser’s logs — filter by level (error/warning/info/log/debug), search the text, or download as text/JSON.

Recording layout (tiled vs solo)

When a browser test runs multiple browsers on one runner, choose how the video is laid out at run-create time (with an account default):

• Tiled (default) — all browsers in one video, arranged in a grid.
• Solo — one browser fills the whole recording surface.

Clean video, no script edits

MaxoPerf suppresses the Chrome “automated test software” infobar and forces the browser headful (so it renders into the recording) without touching your test — your script can keep requesting headless and it will still be captured.

Coverage

Capability	Chrome / Chromium	Firefox
Video	✅	✅
Steps (navigate/click/type/…)	✅	✅
Network HAR	✅	Roadmap (WebDriver BiDi)
Console logs	✅	Roadmap (WebDriver BiDi)

Note

Steps and video are browser-agnostic. Full network/console capture currently uses Chrome’s DevTools; Firefox equivalents (via WebDriver BiDi) are in progress. Playwright tests capture via Playwright’s own tracing/reporter.

Troubleshooting

• No steps/HAR/console, only video — the run probably used a non-browser executor, or a browser engine still on the roadmap. Confirm the test uses Selenium + Chrome.
• Fewer browser rows than virtual users — a test that calls driver.quit() each iteration produces one browser instance per iteration; that’s expected. Capture is bounded per runner to keep overhead low.
• A step seek lands slightly off — seeks are accurate to within a video segment boundary.