How do I group tests ran in a matrix or across multiple runners into the same test run?
By default, each invocation of Playwright is grouped into a test run by a UUID generated when the run begins. To group multiple invocations of Playwright into the same run, set
RECORD_REPLAY_METADATA_TEST_RUN_ID to the same UUID value and that will be used instead of generating a UUID for each.Below is an example which runs three test suites using a matrix in Github Actions but groups the results into the same test run in Replay:
yamljobs: test-run-id: runs-on: ubuntu-latest outputs: testRunId: ${{ steps.testRunId.outputs.testRunId }} steps: - id: testRunId run: echo testRunId=$(npx uuid) >> "$GITHUB_OUTPUT" test: needs: test-run-id runs-on: ubuntu-latest strategy: matrix: product: [frontend, backend, api] steps: - uses: actions/checkout@v3 - uses: actions/setup-node@v3 with: node-version: 16 - name: Install dependencies run: npm ci - name: Run run: npm run test-${{ matrix.product }} env: RECORD_REPLAY_METADATA_TEST_RUN_ID: ${{ needs.test-run-id.outputs.testRunId }}
What’s in the .replay folder?
/.replay/runtimes - this is where the replay browser(s) are installed/ - recorded replays are saved locally in the root folderIf you’re trying to cache the browser download, cache
~/.replay/runtimes or simply cache it before creating replays.How do I confirm that the browsers are installed correctly
Run
ls -alR ~/.replay/runtimes to see the contents of the runtimes folder.What other environment variables can I configure?
REPLAY_SKIP_BROWSER_DOWNLOAD - keeps browsers from being downloaded and installed during npm installRECORD_REPLAY_METADATA_TEST_RUN_TITLE - manually apply a test run title, instead of being inferred from the latest commitRECORD_REPLAY_DIRECTORY - set a custom directory (default: ~/.replay)