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

fix(stats): record noop routes in /v1/routing/stats (#109)

Elyas Mehtabuddin8baf8ec274e7
diff --git a/switchyard/lib/profiles/noop.py b/switchyard/lib/profiles/noop.py
index 70ffc142f..16580589a 100644
--- a/switchyard/lib/profiles/noop.py
+++ b/switchyard/lib/profiles/noop.py
@@ -35,7 +35,7 @@ _NOOP_CONTENT = "pong"
_NOOP_MODEL = "noop"
-def _make_completion(*, request_id: str, created: int) -> ChatCompletion:
+def _make_completion(request_id: str, created: int) -> ChatCompletion:
"""Build the fixed no-op non-streaming completion."""
return ChatCompletion(
id=f"chatcmpl-{request_id}",
@@ -54,7 +54,6 @@ def _make_completion(*, request_id: str, created: int) -> ChatCompletion:
async def _noop_stream(
- *,
request_id: str,
created: int,
) -> AsyncIterator[ChatCompletionChunk]:
@@ -115,14 +114,15 @@ class NoopProfile:
def __init__(
self,
- *,
request_processors: tuple[Any, ...] = (),
response_processors: tuple[Any, ...] = (),
+ stats_accumulator: StatsAccumulator | None = None,
) -> None:
"""Create a no-op profile with a local translation helper."""
self._translation = TranslationEngine()
self._request_processors = request_processors
self._response_processors = response_processors
+ self._stats_accumulator = stats_accumulator
def iter_components(self) -> list[object]:
"""Return lifecycle components in startup order."""
@@ -162,6 +162,7 @@ class NoopProfile:
return NoopProfile(
request_processors=tuple(request_chain),
response_processors=tuple(response_chain),
+ stats_accumulator=stats if enable_stats else None,
)
async def process(self, input: ProfileInput) -> NoopProcessedRequest:
@@ -217,6 +218,9 @@ class NoopProfile:
) -> ChatResponse:
"""Execute no-op response generation with an existing context."""
processed = await self.process_with_context(input, ctx)
+ # Stamp ctx.selected_model so the stats response processor attributes
+ # the response tokens to "noop" instead of "<unknown>".
+ ctx.selected_model = _NOOP_MODEL
openai_request = self._translation.request_to_any_of(
processed.request,
_NOOP_SUPPORTED_TYPES,
@@ -235,6 +239,12 @@ class NoopProfile:
response = ChatResponse.openai_completion(
_make_completion(request_id=request_id, created=created)
)
+ # Count the request only after the response is built, so a translation
+ # failure does not leave a spurious success in /v1/routing/stats. Record
+ # success before rprocess records tokens, which is the order the stats
+ # response processor's record_usage_after_success_attribution expects.
+ if self._stats_accumulator is not None:
+ await self._stats_accumulator.record_success(_NOOP_MODEL)
return await self.rprocess(processed, response)
diff --git a/tests/test_noop_llm_backend.py b/tests/test_noop_llm_backend.py
index 695218129..0a5c4a362 100644
--- a/tests/test_noop_llm_backend.py
+++ b/tests/test_noop_llm_backend.py
@@ -4,11 +4,12 @@
"""Unit tests for the no-op profile."""
from switchyard.lib.profiles import NoopProfile, NoopProfileConfig, ProfileSwitchyard
+from switchyard.lib.stats_accumulator import StatsAccumulator
from switchyard_rust.core import ChatRequest, ChatResponseType, response_type_matches
from switchyard_rust.profiles import ProfileInput
-def _openai_request(*, stream: bool = False) -> ChatRequest:
+def _openai_request(stream: bool = False) -> ChatRequest:
return ChatRequest.openai_chat(
{"model": "noop", "messages": [{"role": "user", "content": "ping"}], "stream": stream}
)
@@ -106,3 +107,22 @@ class TestNoOpProfile:
assert processed.request.model == "noop"
assert processed_response is response
+
+ async def test_noop_call_reports_routing_stats(self) -> None:
+ """The no-op profile self-reports its call and model to the stats accumulator."""
+ stats = StatsAccumulator()
+ profile = NoopProfileConfig().build().with_runtime_components(
+ stats_accumulator=stats
+ )
+
+ await ProfileSwitchyard(profile).call(_openai_request(stream=False))
+
+ snap = stats.snapshot_sync()
+ assert snap["total_requests"] == 1
+ assert "noop" in snap["models"]
+ assert snap["models"]["noop"]["calls"] == 1
+ assert "<unknown>" not in snap["models"]
+ # Response tokens land on "noop", not "<unknown>".
+ assert snap["models"]["noop"]["prompt_tokens"] == 1
+ assert snap["models"]["noop"]["completion_tokens"] == 1
+ assert snap["models"]["noop"]["total_tokens"] == 2