pub struct FleetResolvedRoute {Show 16 fields
pub provider_id: String,
pub provider_exact_id: Option<String>,
pub provider_kind: String,
pub canonical_model: Option<String>,
pub wire_model_id: String,
pub protocol: String,
pub role: Option<String>,
pub loadout: Option<String>,
pub model_class: Option<String>,
pub model_route: Option<String>,
pub reasoning_effort: Option<String>,
pub role_source: Option<String>,
pub loadout_source: Option<String>,
pub model_class_source: Option<String>,
pub model_source: Option<String>,
pub source: String,
}Expand description
Resolved-route detail persisted on a FleetReceipt (#3154).
This is an additive, plain-strings snapshot of the route a fleet worker
resolved to. It deliberately does NOT depend on any codewhale-config route
type so the protocol crate stays free of the route model.
CRITICAL no-secrets invariant: this struct carries ONLY non-sensitive route shape — provider id/kind, model ids, wire protocol, role/loadout/model-class intent, reasoning tier when known, and deterministic intent sources. It must NEVER hold a credential, API key, bearer token, or a base URL that embeds credentials. There is intentionally no field that could carry a secret.
Fields§
§provider_id: StringResolved provider canonical id (e.g. "deepseek").
provider_exact_id: Option<String>Exact configured provider-table id when the worker used one.
This is intentionally additive to provider_id: literal
[providers.custom] resolves to Some("custom"), while the legacy
idless root custom route resolves to None. Keeping the distinction
prevents a receipt from silently collapsing two different credential
and endpoint authorities into the same generic custom label.
provider_kind: StringResolved provider kind (e.g. "deepseek").
canonical_model: Option<String>Canonical, provider-agnostic model identity, when known.
wire_model_id: StringProvider-owned wire model id placed on the request.
protocol: StringSelected wire protocol (e.g. "chat_completions").
role: Option<String>Effective Fleet role intent, when one applied.
loadout: Option<String>Effective Fleet loadout intent, when one applied.
model_class: Option<String>Original task-level model-class intent, when authored separately from
loadout. Profile model_class_hint is normalized into loadout.
model_route: Option<String>Runtime model-route seam used by sub-agent routing (inherit, faster,
auto, or fixed).
reasoning_effort: Option<String>Concrete reasoning tier, when it is known by the route resolver path.
role_source: Option<String>Deterministic source for the effective role intent.
loadout_source: Option<String>Deterministic source for the effective loadout intent.
model_class_source: Option<String>Deterministic source for the model-class hint, when present.
model_source: Option<String>Deterministic source for the model selector used by the resolver.
source: StringHow the route was produced (e.g. "resolver").
Trait Implementations§
Source§impl Clone for FleetResolvedRoute
impl Clone for FleetResolvedRoute
Source§fn clone(&self) -> FleetResolvedRoute
fn clone(&self) -> FleetResolvedRoute
1.0.0 (const: unstable) · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
source. Read more