Skip to content
Error boundary / 500 page

Error boundary / 500 page

The failure surface for an unexpected server or client crash, which must not lose the user's work or their trust.

System & Edge States · 8 checks · 5 resources

Before you ship

8
  1. 01Plain-language statement that the failure is ours, not the user's input
  2. 02Retry action that re-attempts only the failed unit — the component or request, not a full page reload where a boundary can recover locally
  3. 03Error reference identifier displayed and copyable — the same ID present in server logs, so support can trace it
  4. 04Escape routes — links to the previous page, the dashboard, and the status page
  5. 05Scoped boundary — a crash in one panel does not blank the entire application shell
  6. 06Unsaved state preserved where the failure is client-side — draft content retained in local storage and offered on recovery
  7. 07Automatic report with an opt-in description field — telemetry sent without the user having to file a ticket
  8. 08No stack traces, internal hostnames or SQL in the user-facing message

Resources that help you build this

5