Skip to main content
Something broke on our side that we didn’t anticipate. Every internal_error is a bug. The requestId field is the link between you and our logs.

What triggers it

  • Unhandled exceptions on a code path we haven’t covered.
  • Database / Redis connectivity loss.
  • Storage layer (S3) refused the call for a reason we can’t classify.
  • A feature flag is in an inconsistent state.
If a third-party platform is down, you’ll see platform_unavailable instead — internal_error is exclusively about us.

Response shape

internal_error.json
The message field is intentionally generic — internal errors don’t leak server-side details. The requestId is the way to reference this specific failure.

How to handle

  1. Retry once with the same Idempotency-Key.
  2. If it fails again, file an issue or open a support thread with the requestId.
  3. We do not return server stack traces; please don’t try to parse message for diagnosis.