Struct MetricsCallback 

pub struct MetricsCallback { /* private fields */ }

Callback that collects latency, token, and error metrics.

Uses the standard RunEvent lifecycle events to measure model and tool call latency, accumulate token usage, and count errors.

Call snapshot() at any time to read the current metrics, or reset() to zero them out.

Implementations

impl MetricsCallback

pub fn new() -> Self

pub async fn snapshot(&self) -> MetricsSnapshot

Take a snapshot of the current metrics.

pub async fn record_tokens(&self, input_tokens: u64, output_tokens: u64)

Record token usage externally (e.g. from a ChatResponse).

This allows callers that have access to the actual TokenUsage from model responses to feed it into the metrics.

pub async fn reset(&self)

Reset all metrics.

Trait Implementations

impl CallbackHandler for MetricsCallback

fn on_event<'life0, 'async_trait>( &'life0 self, event: RunEvent, ) -> Pin<Box<dyn Future<Output = Result<(), SynapticError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

impl Default for MetricsCallback

fn default() -> Self

Returns the “default value” for a type.