Error Codes

HTTP error codes returned by the Metrion proxy.

Most errors from the proxy come directly from the AI provider and are forwarded as-is. Metrion adds a small set of its own error responses for authentication, rate limiting, and timeout cases.

Metrion-specific errors

Status	Meaning
401 Unauthorized	Missing or invalid Metrion token; or your provider key is not configured for stored mode.
429 Too Many Requests	Rate limit exceeded (100 requests per minute per token), or you have reached your monthly limit (Free plan: 10,000 requests per month).
504 Gateway Timeout	The provider did not respond within 30 seconds.
400 Bad Request	Invalid request body — typically a missing model, messages, or max_tokens field.
403 Forbidden	The feature you are accessing is not available on your current plan.
502 Bad Gateway	The provider returned an unexpected error that Metrion could not parse.

Provider errors

Errors originating from the AI provider (4xx and 5xx) are forwarded to you exactly as the provider returned them, including the original HTTP status code and error body. Metrion does not modify provider error responses.

For example, if Anthropic returns a 400 because your max_tokens exceeds the model’s context window, you will receive the exact same 400 response body that Anthropic returned.

Viewing errors in the dashboard

Open the Errors tab in your dashboard to see a breakdown of error codes over time, including a timeline chart and a table of your 50 most recent errors.

Error messages from both Metrion and the provider are captured and displayed on the Logs page, so you can inspect individual failed requests.