Name: Selenium & Playwright Interview Questions
Availability: InStock

Ten questions in the style of a technical screen: each one has a plausible wrong answer that interviewers hear often.

1. A Selenium test sets an implicit wait of 10 seconds and also uses a 10-second WebDriverWait. What does the Selenium documentation say about this?

It is recommended: the implicit wait covers finding, the explicit wait covers state. Avoid it: mixing the two can produce unpredictable wait times. The explicit wait silently disables the implicit wait.

Mixing implicit and explicit waits can stack timeouts in unexpected ways. Pick explicit waits and leave the implicit wait at zero.

2. Your team runs 200 tests one after another with a single-threaded runner. You point them at a Selenium Grid with 50 free slots. What happens to the run time?

It drops to roughly a fiftieth. It drops, but only by half because of network overhead. Roughly nothing changes: the runner still starts one session at a time.

Grid provides capacity; parallelism comes from the test runner. Without parallel execution configured, only one slot is ever used.

3. Which Playwright line is likely to be flaky when the cart count updates asynchronously?

expect(await cart.textContent()).toBe('3') await expect(cart).toHaveText('3') Both are equally reliable because Playwright auto-waits.

The first form reads the text once and compares it. Only the web-first assertion retries until the text matches or the timeout expires.

4. A Playwright click fails with a strict mode violation. What does that mean?

The element was not visible or enabled before the timeout. The locator matched more than one element, so Playwright refused to guess. The page violated a Content Security Policy.

Locators are strict: an action on a locator that resolves to several elements throws. Narrow it by scoping to a container or filtering, rather than reaching for .first() by reflex.

5. You shard a Playwright suite across four CI jobs. How do you get one HTML report?

Each shard writes the HTML report into the same folder; the last one wins. Sharding cannot produce a combined report. Use the blob reporter on each shard, then npx playwright merge-reports.

Blob reports contain all the run information needed to merge. A final job collects them into one folder and merges them into an HTML report.

6. In Cypress, why does const el = cy.get('button'); el.click() not behave like Selenium code?

Cypress commands are queued and yield their subject later; cy.get does not return the element. cy.get returns a Promise that must be awaited. Cypress only allows one variable per test.

Commands are enqueued and run later; you work with yielded values in .then() or through aliases. Cypress chains are not Promises, and await breaks the queue.

7. A Cypress 16 upgrade breaks tests that call Cypress.env('apiKey'). What is the right replacement for a secret?

Cypress.expose('apiKey') cy.env(['apiKey']).then(({ apiKey }) => ...) cy.exec('echo $API_KEY')

cy.env() only hands over the values you ask for and keeps secrets out of browser state. Cypress.expose() is for public values, and cy.exec() was removed in Cypress 16.

8. After moving to Appium 3, a test that used --allow-insecure=adb_shell can no longer run shell commands. What fixed it?

Downgrading Node.js to 18. Adding appium:adbShell: true to the capabilities. Scoping the flag to the driver: --allow-insecure=uiautomator2:adb_shell.

Appium 3 requires insecure feature flags to carry a driver prefix. It also needs Node.js 20.19 or newer, so downgrading Node would stop the server from starting.

9. Why is XPath a last-resort locator in Appium?

The driver must serialise the whole UI hierarchy to XML and evaluate against it, which is slow on complex screens. Appium 3 removed XPath support. XPath only works on Android.

XPath still works on both platforms, but accessibility IDs and native queries (UiAutomator, predicate strings, class chains) avoid the expensive hierarchy dump.

10. A TestNG suite uses parallel="methods" and a static WebDriver driver in a base class. What do you expect?

Everything works, but each test is slower. Tests act on and quit each other's browsers, failing at random. TestNG detects the static field and runs the methods sequentially.

All threads share one field, so they overwrite and quit each other's sessions. Keep the driver in a ThreadLocal or create it per test method.

For hands-on practice with the frameworks behind these questions, continue with the Playwright course or Selenium with TestNG; mobile candidates can follow the Appium with WebdriverIO course.

Knowledge check

You finished Selenium & Playwright Interview Questions