Struct RequestSample 

pub struct RequestSample {
    pub index: usize,
    pub worker: usize,
    pub start_s: f64,
    pub end_s: f64,
    pub token_times_s: Vec<f64>,
    pub generated_tokens: u32,
    pub prompt_tokens: u32,
    pub outcome: Outcome,
    pub in_flight_at_start: usize,
    pub drained: bool,
}

One retained per-request sample (§4.4.5). Every time is an offset in seconds from a single band-wide origin, so the wall-clock span is computable across workers without reconstructing anything.

Fields

index: usize

Monotone index in issue order.

worker: usize

Which closed-loop worker issued it.

start_s: f64

Offset of the request start from the band origin.

end_s: f64

Offset of the request’s completion (or abandonment) from the band origin.

token_times_s: Vec<f64>

Arrival offsets of each streamed token. Empty when not streaming.

generated_tokens: u32

Generated (completion) tokens, counted per the receipt’s tokenization block.

prompt_tokens: u32

Prompt tokens, when the server reports them.

outcome: Outcome

How the request ended.

in_flight_at_start: usize

Concurrent requests in flight at the instant this one was issued. The direct, per-request evidence that the client was actually concurrent.

drained: bool

True when this request completed after the measurement window closed, i.e. during the §4.4.7 drain.

Implementations

impl RequestSample

pub fn ttft_ms(&self) -> Option<f64>

§4.4.3 ttft_ms — request start to first token byte at the client. None when no token was ever observed.

pub fn decode_tok_s(&self) -> Option<f64>

§4.4.3 decode_tok_s for this request: (generated tokens - 1) / (last token time - first token time).

None when fewer than two tokens were observed, where the quantity is undefined rather than zero.

pub fn itl_gaps_ms(&self) -> Vec<f64>

§4.4.3 itl_ms — this request’s inter-token gaps, for pooling.

pub fn counts_toward_aggregate(&self) -> bool

True when this sample contributes to agg_tok_s’s numerator (§4.4.3: completed and non-truncated).

Trait Implementations

impl Clone for RequestSample

fn clone(&self) -> Self

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for RequestSample

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for RequestSample

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for RequestSample

fn eq(&self, other: &Self) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl Serialize for RequestSample

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for RequestSample