Enum Error 

#[non_exhaustive]
pub enum Error {
    Connection {
        message: String,
        kind: TransportErrorKind,
        source: Option<Box<dyn Error + Send + Sync>>,
    },
    Request {
        message: String,
        code: Option<String>,
        status: u16,
        request: Option<Box<RequestMetadata>>,
    },
    Server {
        message: String,
        code: Option<String>,
        status: u16,
        request: Option<Box<RequestMetadata>>,
    },
    InputTooLong {
        message: String,
        model: Option<String>,
        request: Option<Box<RequestMetadata>>,
    },
    ModelLoadFailed {
        message: String,
        model: Option<String>,
        error_class: ModelLoadErrorClass,
        permanent: bool,
        attempts: u32,
        request: Option<Box<RequestMetadata>>,
    },
    ResourceExhausted {
        message: String,
        model: Option<String>,
        retries: u32,
        request: Option<Box<RequestMetadata>>,
    },
    EstimateUnroutable {
        message: String,
        code: Option<String>,
        request: Option<Box<RequestMetadata>>,
    },
    Provisioning {
        message: String,
        gpu: Option<String>,
        retry_after: Option<Duration>,
    },
    ModelLoading {
        message: String,
        model: Option<String>,
    },
    LoraLoading {
        message: String,
        lora: Option<String>,
        model: Option<String>,
    },
    Pool {
        message: String,
        pool_name: Option<String>,
        state: Option<String>,
    },
    Decode(String),
    InvalidRequest(String),
    Io(Error),
}

Everything that can go wrong talking to a SIE server.

The variant fields are the failure’s evidence: message is what the server said, code is its error code, status the HTTP status, and request the metering metadata that survived. Read them through the accessors when you only need one.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

Connection

The request never produced an HTTP response.

Fields

message: String

kind: TransportErrorKind

source: Option<Box<dyn Error + Send + Sync>>

Request

A 4xx the SDK does not model more specifically.

Fields

message: String

code: Option<String>

status: u16

request: Option<Box<RequestMetadata>>

Server

A 5xx the SDK does not model more specifically.

Fields

message: String

code: Option<String>

status: u16

request: Option<Box<RequestMetadata>>

InputTooLong

The input exceeded the model’s context window (400 INPUT_TOO_LONG).

Fields

message: String

model: Option<String>

request: Option<Box<RequestMetadata>>

ModelLoadFailed

The worker could not load the model (502 MODEL_LOAD_FAILED).

Fields

message: String

model: Option<String>

error_class: ModelLoadErrorClass

permanent: bool

false only for cooldown-suppressed OOM/network failures, which may succeed later.

attempts: u32

request: Option<Box<RequestMetadata>>

ResourceExhausted

The server ran out of device memory and the SDK exhausted its retries.

Fields

message: String

model: Option<String>

retries: u32

request: Option<Box<RequestMetadata>>

EstimateUnroutable

/v1/estimate could not route the request to a rate identity.

Fields

message: String

code: Option<String>

request: Option<Box<RequestMetadata>>

Provisioning

No capacity, and either the caller opted out of waiting or the budget ran out.

Fields

message: String

gpu: Option<String>

retry_after: Option<Duration>

ModelLoading

The model was still loading when the provision budget expired.

Fields

message: String

model: Option<String>

LoraLoading

A LoRA adapter was still loading after the retry cap.

Fields

message: String

lora: Option<String>

model: Option<String>

Pool

A pool operation failed.

Fields

message: String

pool_name: Option<String>

state: Option<String>

Decode(String)

A response body could not be decoded as the documented shape.

InvalidRequest(String)

The caller supplied arguments the SDK rejects before sending anything.

Io(Error)

Local I/O failed (reading a file to upload, for example).

Implementations

impl Error

pub fn status(&self) -> Option<u16>

HTTP status that produced this error, when there was a response.

pub fn code(&self) -> Option<&str>

Server error code (X-SIE-Error-Code or the body’s code field).

pub fn request_metadata(&self) -> Option<&RequestMetadata>

Metadata the server attached to the failing request, when any survived parsing.

pub fn retry_after(&self) -> Option<Duration>

The server’s Retry-After hint, when it sent one and the SDK gave up anyway.

pub fn is_server_error(&self) -> bool

True for every failure the server attributed to itself (5xx family).

pub fn is_request_error(&self) -> bool

True for every failure the server attributed to the request (4xx family).

pub fn is_capacity_error(&self) -> bool

True when the SDK gave up waiting for capacity rather than hitting a hard failure.

Trait Implementations

impl Debug for Error

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Display for Error

fn fmt(&self, __formatter: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Error for Error

fn source(&self) -> Option<&(dyn Error + 'static)>

Returns the lower-level source of this error, if any.

fn description(&self) -> &str

👎Deprecated since 1.42.0:

use the Display impl or to_string()

fn cause(&self) -> Option<&dyn Error>

👎Deprecated since 1.33.0:

replaced by Error::source, which can support downcasting

fn provide<'a>(&'a self, request: &mut Request<'a>)

🔬This is a nightly-only experimental API. (error_generic_member_access)

Provides type-based access to context intended for error reports.

impl From<Error> for Error

fn from(source: Error) -> Self

Converts to this type from the input type.