InferNode
| HTTP | type | code | Action |
|---|
| 400 | invalid_request_error | parameter_not_supported | Field missing, out of range or not supported by that model (param names it). |
| 400 | invalid_request_error | provider_invalid_request | The prompt plus max_tokens exceed the model's context_window; the message includes the count. Lower max_tokens or shorten the prompt (Context exceeded). |
| 401 | authentication_error | identity.invalid_token | The Authorization header is missing or does not have the form Bearer <llave>. |
| 401 | authentication_error | identity.unauthorized | The key does not exist, was revoked, or the account or organization is suspended. Generate a new one or contact the administrator. |
| 402 | billing_error | insufficient_credits | Out of credits. Contact the administrator. |
| 403 | permission_error | | The key does not have access to the model or resource. |
| 404 | not_found_error | model_not_found | Use a name from GET /v1/models. |
| 429 | rate_limit_error | rate_limit_exceeded | Wait Retry-After seconds; details.dimension indicates the counter (rpm, rpd, tpm, tpd, ipm). |
| 502 | provider_error | | The provider failed. Retry with exponential backoff. |
| 503 | provider_error | provider_circuit_open | That model failed several times in a row and InferNode stopped sending it requests for a while; the message says how long to wait (retry after 30s). Retry after that time or switch models. |
| 504 | provider_error | provider_timeout | The model did not respond in time; the gateway is still up. Switch models or retry later (InferNode does not respond). |
| 503 | maintenance_error | | Maintenance. Retry later. |
Every error response includes request_id; include it when you report the problem.
Batches
When you upload the file or manage the batch (Process requests in batches):
| HTTP | type | code | Action |
|---|
| 400 | invalid_request_error | file_purpose_invalid | Upload with purpose=batch, the only value it accepts, and create the batch with a file uploaded that way. |
| 400 | invalid_request_error | invalid_file_format | The file does not start with JSON. Check that it is JSONL. |
| 400 | invalid_request_error | file_in_use | You are deleting an input file that a live batch uses. Wait for the batch to finish. |
| 400 | invalid_request_error | storage_quota_exceeded | Out of storage space. Delete files you no longer use. |
| 400 | invalid_request_error | batch_not_cancellable | The batch is already in finalizing or has finished. |
| 404 | not_found_error | file_not_found / batch_not_found | The identifier does not exist, belongs to another organization or was deleted. |
| 413 | invalid_request_error | file_too_large | The file is larger than 100 MiB. Split it into several batches. |
| 429 | rate_limit_error | too_many_active_batches | You already have as many batches running as the plan allows. Wait for them to finish. |
| 503 | internal_error | storage_unavailable | Storage did not respond. Retry. |
When the problem is with the whole file, the batch ends in failed and errors.data lists up to 100 causes: invalid_json_line, duplicate_custom_id, empty_file, too_many_tasks, url_mismatch, model_mismatch, model_not_found, model_not_batchable or enqueued_request_limit_exceeded.
When the problem is with a single line, that line appears in the error file with stream_not_supported, insufficient_credits, spending_cap_exceeded, authorization_revoked, model_not_found, batch_cancelled or batch_expired. Lines with batch_cancelled and batch_expired are not charged.
Agents and workflows
| HTTP | code | Action |
|---|
| 400 | invalid_query, invalid_body | Fix the parameters or the body. |
| 401 | unauthorized | Sign in again or renew the token. |
| 404 | agent_not_found, workflow_not_found | It does not exist, is not yours or was deleted. |
| 409 | agent_in_use | The agent is connected to a workflow; disconnect it before deleting. |
| 409 | unpublished_changes | Publish before running. |
| 422 | channel_incomplete | A channel is not configured; check the Canales (Channels) tab. |
Labs (Frida)
| HTTP | code | Action |
|---|
| 400 | LAB_INVALID_RESOURCES | Malformed quantity (16Gi, 500m, 24576) or size and resources at the same time. |
| 403 | | Only the owner can delete or share. |
| 404 | | It does not exist or belongs to another Client. |
| 409 | | Duplicate name or deletion in progress. |
| 409 | LAB_RESIZE_REQUIRES_STOP | The change does not fit against what is running; details names the axis and the notebook. Repeat with stopRunningWorkloads: true if you accept stopping it. If the message says that stopping would not be enough, first free what is taking up the budget (a pipeline run). |
| 409 | LAB_CAPACITY_UNAVAILABLE | When resizing up: no capacity right now. Wait or request less. |
| 409 | LAB_RESIZE_NOT_READY | The lab is being created or deleted. Wait for ready. |
| 409 | LAB_RESIZE_STOP_TIMED_OUT | The notebooks were stopped but usage did not drop in time; nothing was applied. Retry. |
| 412 | LAB_RESIZE_STALE_VERSION | The If-Match is stale: read the lab again and repeat. |
| 422 | LAB_POLICY_UNAVAILABLE | Size not offered or resource above the maximum; details indicates the resource, the request and the limit. |
| 422 | LAB_CAPACITY_UNAVAILABLE | When creating: no capacity right now. Wait or request a smaller size. |
| 422 | LAB_CEILING_EXCEEDED | When resizing with custom resources: an axis exceeds the plan's ceiling; details includes the axis, the request and the limit. |
| 422 | LAB_MEMBER_NOT_IN_ORGANIZATION, LAB_MEMBER_IS_OWNER, LAB_INVALID_ROLE | When sharing: the person is not in the Client, is the owner, or the role is not editor. |
Messages in Studio
| Message | Page |
|---|
| "No tienes permisos para ver los accesos de este cliente." (You do not have permission to view this client's access.) | No permission |
| "Tu laboratorio rechazó una petición" (Your lab rejected a request) | The notebook does not start |
| A volume's magnifying glass never opens | The volume file browser does not open |
| The notebook starts empty after a restart | The notebook lost its files on restart |
| "No hay capacidad disponible ahora mismo para esta talla o recursos" (There is no capacity available right now for this size or resources) | Exceeds the quota |
| "Hay conexiones sin configuración completa." (Some connections are not fully configured.) | Publish a workflow |
| "El proveedor rechazó las credenciales." (The provider rejected the credentials.) | Shared credentials |
| "Ese dominio no tiene inicio de sesión único configurado." (That domain does not have single sign-on configured.) | SSO |
| "Este enlace no es válido o ya expiró." (This link is not valid or has expired.) | Accessing Studio |
| "No pudimos conectar con el proveedor de identidad de tu organización." (We could not connect to your organization's identity provider.) | SSO does not let you in |
| "El agente no pudo responder." (The agent could not respond.) / "El agente tardó demasiado en responder." (The agent took too long to respond.) | The agent does not respond |
| "El modelo guardado ya no está disponible para este Agent." (The saved model is no longer available for this Agent.) | Model not available |
| "Panel inicializándose" (Dashboard initializing) / "Snapshot antiguo" (Stale snapshot) / "sin telemetría" (no telemetry) | The Panel shows no data |
