playwright-trace-analysis

Provide the path to the trace.zip and the analysis starts with an overview of trace steps, network activity, and screenshots. It identifies the first meaningful failure, corroborates with screenshots or network logs, and reports a root cause with cited evidence.

Compare traces across retries: a consistent failure at the same step indicates a deterministic app or test bug, while pass-on-retry with timing-sensitive signals like loading spinners suggests flakiness. Multiple traces failing identically should not be classified as flaky.

Manually created traces using the tracing.start/stop API may not be recognized by automated tools. Verify the zip contains trace.trace and trace.network files, then parse the newline-delimited JSON directly with the provided Python scripts.

Application code often logs critical errors via console.log instead of console.error, so severity-only filtering misses them. Scan console messages at all levels for failure keywords like TypeError, fetch failed, ECONNREFUSED, and timeout.

Yes, with caveats. Electron traces capture renderer-process console and network activity, but main-process API calls are not recorded. Watch for IPC errors, preload script failures, webview loading issues, and multi-window confusion.