> ## Documentation Index
> Fetch the complete documentation index at: https://docs.nebius.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Errors in the REST API

When a request fails, the Nebius AI Cloud API returns a status code and an error message. For example:

```bash theme={null}
curl -v -XPOST https://api.nebius.cloud/compute/v1/instances
...
* Request completely sent off
< HTTP/2 400
...
<
* Connection #0 to host api.nebius.cloud left intact
{"code":3, "message":"invalid resource size type: <nil>"}%
```

The response header contains an HTTP status code, and the response body includes the following:

* `code`: Numeric canonical gRPC status code that correlates with the HTTP status code based on the [Google RPC specification](https://github.com/googleapis/googleapis/blob/master/google/rpc/code.proto).
* `message`: Description of the error.

In this example, the gRPC status code is `3 INVALID_ARGUMENT`, which corresponds to `400 Bad Request` in HTTP.

The Nebius AI Cloud API only uses canonical codes in the [full list of gRPC status codes](https://grpc.io/docs/guides/status-codes/#the-full-list-of-status-codes). The following table describes the possible status codes:

<div className="first-expanded-column">
  | gRPC status code | HTTP status code | Description |
  | - | - | - |
  | `0 OK` | `200 OK` | The request completed successfully. |
  | `1 CANCELLED` | `499 Client Closed Request` | You canceled the request. |
  | `3 INVALID_ARGUMENT` | `400 Bad Request` | The request is invalid or is out of the service constraints. Make sure that you specified the method, headers and request body correctly. |
  | `4 DEADLINE_EXCEEDED` | `504 Gateway Timeout` | The request timed out.<br /><br />For requests that change the state of the system, this error may occur even if the request completed successfully but the response was delayed. |
  | `5 NOT_FOUND` | `404 Not Found` | The requested resource doesn't exist. For example, it may have been deleted. Make sure that you address an existing resource — check its ID or name in the request. |
  | `6 ALREADY_EXISTS` | `409 Conflict` | The resource with the specified parameters already exists. For example, you try to create a virtual machine with the name that another virtual machine already has. To fix the error, use a different name. |
  | `7 PERMISSION_DENIED` | `403 Forbidden` | You don't have the permission to perform the request. Check that you have the required IAM role.<br /><br />This error is different from `16 UNAUTHENTICATED` because `7 PERMISSION_DENIED` indicates a missing role while `16 UNAUTHENTICATED` points to the incorrect credentials. |
  | `8 RESOURCE_EXHAUSTED` | `429 Too Many Requests` | The [quotas](/overview/quotas) of the requested resources are exhausted. Clean up the resources or increase the quotas.<br /><br />If the request tries to exceed the service constraints, another error `3 INVALID_ARGUMENT` occurs. |
  | `9 FAILED_PRECONDITION` | `400 Bad Request` | The request can't be performed because of the current state of resources. For example, the service only allows deleting empty directories but you try to delete a non-empty one. In that case, first delete resources within the directory and then delete the directory itself. |
  | `10 ABORTED` | `409 Conflict` | The request is aborted, usually due to a concurrency issue such as a resource version mismatch. |
  | `11 OUT_OF_RANGE` | `400 Bad Request` | The request attempts to use a value outside the valid parameter range. Change the value so it's within the range. |
  | `13 INTERNAL` | `500 Internal Server Error` | An important system condition isn't met. |
  | `14 UNAVAILABLE` | `503 Service Unavailable` | The service is unavailable. Try again later. |
  | `16 UNAUTHENTICATED` | `401 Unauthorized` | The request is sent with invalid authentication credentials. Check your credentials.<br /><br />This error is different from `7 PERMISSION_DENIED` because `16 UNAUTHENTICATED` only checks the credentials, not roles. |
</div>

The list doesn't include the `2 UNKNOWN`, `12 UNIMPLEMENTED` and `15 DATA_LOSS` errors because the REST API by Nebius AI Cloud never returns them.