[Question] How to increase page lookup timing

Migrating from puppeteer to playwright I have noticed an increase in the time to load a simple page using page.goto. Take the example below…

// testing.test.ts
import { test, expect } from '@playwright/test';

test('Testing page.goto', async ({ page }) => {
  await page.goto('https://google.com');
  expect(true).toBe(true);
});

The html reporting shows it takes ~2s for the goto step to complete. Comparing this to puppeteer times on the order of 300ms.

To my understanding, a new page is created for each test and seems like a necessary tradeoff to enable parallelism. That said I am curious if this time is typical or if there is something I can configure or even a custom fixture, to increase this speed?

My full playwright config

// playwright.config.ts
import { PlaywrightTestConfig } from '@playwright/test';

const isCI = process.env.CI === 'true';

const config: PlaywrightTestConfig = {
  use: {
    headless: true,
    locale: 'en-us',
    viewport: { width: 785, height: 600 },
    trace: 'off',
    screenshot: 'off', // already testing screenshots
    video: 'retain-on-failure',
    launchOptions: {
      ignoreDefaultArgs: ['--hide-scrollbars'],
      args: ['--use-gl=egl'],
    },
  },
  reporter: [['html', { open: 'never', outputFolder: 'html_report' }], isCI ? ['line'] : ['list']],
  expect: {
    toMatchSnapshot: {
      threshold: 0,
      maxDiffPixels: 4,
    },
  },
  forbidOnly: isCI,
  timeout: 10 * 1000,
  preserveOutput: 'failures-only',
  snapshotDir: 'screenshots',
  testDir: 'tests',
  outputDir: 'test_failures',
  fullyParallel: true,
  projects: [
    {
      name: 'chrome',
      use: {
        browserName: 'chromium',
      },
    },
  ],
};

[dgozman]

Good to know, I’ll take a look at using a single page between tests. So if this context creation is not for parallelism, I should be able to reuse a single page across tests even with fullyParallel enabled, correct?

Not really. If you mark your suite as serial, tests in that suite will run in order. There is no way to reuse the page between tests that literally run in parallel at the same time 😄 If you manually split your tests into files/suites, and only reuse the page inside each suite, you’ll have full control of what can run in parallel.

[dgozman]

Closing since there doesn’t seem to be any action item here.