Observability

Live
Trace requests without leaking credentials or sensitive payloads.
Open in ChatGPT
(opens in new tab)
Last verified: 2026-08-11

Record the right fields

Capture internal job ID, ModelRush request/generation ID, model, routed region, status, latency, token or media usage, attempt number, and normalized error code.

Protect payloads

Redact Authorization headers and avoid logging full prompts, uploaded media, or transcription content by default. Use hashes or controlled samples when debugging requires correlation.

Dashboard workflow

Use Requests to inspect ModelRush-side status and compare it with your application trace. A shared request ID shortens support investigations.