Test isolation and parallelism

Test user-visible behavior rather than implementation details such as function names or CSS classes, because end users interact only with the rendered output. Make each test fully isolated: every test runs independently with its own local storage, session storage, data, and cookies, which improves reproducibility, eases debugging, and prevents cascading failures. Use beforeEach hooks for repeated setup such as navigating or signing in so no test relies on another, though a little duplication is acceptable when it keeps simple tests clearer; the signed-in state can also be reused across tests with a setup project. Playwright runs tests in parallel by default, while tests in a single file run in order in the same worker process; independent tests in one file can run in parallel via test.describe.configure({ mode: 'parallel' }), and the suite can be sharded across multiple machines with --shard=1/3. Scope: Playwright. Reviewed October 4, 2026.