Skip to content

Switchyard lets LLM applications route traffic across models and providers while preserving native OpenAI and Anthropic API compatibility - enabling flexible model selection, benchmarking, and cost/performance optimization.

Rust

feat(switchyard-server): `/v1/stats/reset` (#213)

Graham Kinga6a4dcea7a98
diff --git a/crates/switchyard-server/src/lib.rs b/crates/switchyard-server/src/lib.rs
index eb0a8a898..e97de8f12 100644
--- a/crates/switchyard-server/src/lib.rs
+++ b/crates/switchyard-server/src/lib.rs
@@ -300,6 +300,7 @@ pub fn build_switchyard_router(state: ServerState) -> Router {
.route("/v1/messages/count_tokens", post(anthropic_count_tokens))
.route("/v1/models", get(models))
.route("/v1/stats", get(get_stats))
+ .route("/v1/stats/reset", post(reset_stats))
.route("/metrics", get(prometheus_metrics))
.route("/health", get(health))
.fallback(not_found)
@@ -766,6 +767,11 @@ async fn get_stats(State(state): State<ServerState>) -> Json<StatsSnapshot> {
Json(state.stats.snapshot())
}
+async fn reset_stats(State(state): State<ServerState>) -> Json<Value> {
+ state.stats.reset();
+ Json(json!({"status": "reset"}))
+}
+
async fn health() -> Json<Value> {
Json(json!({"status": "ok"}))
}
diff --git a/crates/switchyard-server/src/stats/accumulator.rs b/crates/switchyard-server/src/stats/accumulator.rs
index 2aa924ec7..28e3f80e6 100644
--- a/crates/switchyard-server/src/stats/accumulator.rs
+++ b/crates/switchyard-server/src/stats/accumulator.rs
@@ -134,6 +134,11 @@ impl StatsAccumulator {
inner.snapshot()
}
+ /// Clears all accumulated stats.
+ pub(crate) fn reset(&self) {
+ *self.lock() = StatsAccumulatorInner::default();
+ }
+
fn lock(&self) -> MutexGuard<'_, StatsAccumulatorInner> {
self.inner.lock()
}
@@ -562,6 +567,24 @@ mod tests {
);
}
+ #[test]
+ fn reset_clears_backend_classifier_and_cache_eligibility_state() {
+ let stats = StatsAccumulator::default();
+ stats.record_success("model/a", 10.0, Some("strong"));
+ stats.record_usage("model/a", usage(10, 5), 15.0, Some("strong"));
+ stats.record_routing_overhead(5.0);
+ stats.record_classifier_success("model/classifier", Some(usage(4, 1)), 2.0);
+ let probe = prefix_probe(&json!({
+ "messages": [{"role": "user", "content": "repeat me"}],
+ }));
+ stats.prefix_eligibility("model/a", &probe);
+
+ stats.reset();
+
+ assert_eq!(stats.snapshot(), StatsSnapshot::default());
+ assert_eq!(stats.prefix_eligibility("model/a", &probe), 0.0);
+ }
+
#[test]
fn theoretical_cache_hit_rate_is_switch_aware() {
let stats = StatsAccumulator::default();
diff --git a/crates/switchyard-server/tests/server.rs b/crates/switchyard-server/tests/server.rs
index 0581ee153..69ed372ca 100644
--- a/crates/switchyard-server/tests/server.rs
+++ b/crates/switchyard-server/tests/server.rs
@@ -296,6 +296,40 @@ async fn stats_accumulates_buffered_success_error_and_shared_routes() -> TestRes
Ok(())
}
+#[tokio::test]
+async fn stats_reset_returns_confirmation_and_clears_all_stats() -> TestResult {
+ let (_upstream, app) = test_app(&[(ROUTE_MODEL, &["model/a"])]).await?;
+ assert_eq!(
+ send(
+ &app,
+ "POST",
+ "/v1/chat/completions",
+ Some(json!({
+ "model": ROUTE_MODEL,
+ "messages": [{"role": "user", "content": "hello"}]
+ })),
+ )
+ .await?
+ .status,
+ StatusCode::OK
+ );
+
+ let reset = send(&app, "POST", "/v1/stats/reset", None).await?;
+ assert_eq!(reset.status, StatusCode::OK);
+ assert_eq!(reset.json()?, json!({"status": "reset"}));
+
+ let stats = send(&app, "GET", "/v1/stats", None).await?.json()?;
+ assert_eq!(stats["total_requests"], 0);
+ assert_eq!(stats["total_errors"], 0);
+ assert_eq!(stats["total_tokens"], empty_token_totals());
+ assert_eq!(stats["models"], json!({}));
+ assert_eq!(stats["tiers"], json!({}));
+ assert_eq!(stats["routing_overhead"]["count"], 0);
+ assert_eq!(stats["classifier"]["total_requests"], 0);
+ assert_eq!(stats["classifier"]["models"], json!({}));
+ Ok(())
+}
+
#[tokio::test]
async fn metrics_exposes_switchyard_otel_instruments() -> TestResult {
const MODEL: &str = "model/metrics-buffered";