biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Convert formatting to use ruff instead of yapf + isort (#26247)

Browse Source 
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2025-10-05 15:06:22 +01:00
committed by GitHub
parent 17edd8a807
commit d6953beb91
1508 changed files with 115244 additions and 94146 deletions
Download Patch File Download Diff File Expand all files Collapse all files

44
vllm/tracing.py
View File

@@ -17,12 +17,15 @@ otel_import_error_traceback: Optional[str] = None
try:
from opentelemetry.context.context import Context
from opentelemetry.sdk.environment_variables import (
OTEL_EXPORTER_OTLP_TRACES_PROTOCOL)
OTEL_EXPORTER_OTLP_TRACES_PROTOCOL,
)
from opentelemetry.sdk.trace import TracerProvider
from opentelemetry.sdk.trace.export import BatchSpanProcessor
from opentelemetry.trace import SpanKind, Tracer, set_tracer_provider
from opentelemetry.trace.propagation.tracecontext import (
TraceContextTextMapPropagator)
TraceContextTextMapPropagator,
)
_is_otel_imported = True
except ImportError:
# Capture and format traceback to provide detailed context for the import
@@ -30,6 +33,7 @@ except ImportError:
# memory leaks.
# See https://github.com/vllm-project/vllm/pull/7266#discussion_r1707395458
import traceback
otel_import_error_traceback = traceback.format_exc()
class Context: # type: ignore
@@ -49,13 +53,15 @@ def is_otel_available() -> bool:
return _is_otel_imported
def init_tracer(instrumenting_module_name: str,
otlp_traces_endpoint: str) -> Optional[Tracer]:
def init_tracer(
instrumenting_module_name: str, otlp_traces_endpoint: str
) -> Optional[Tracer]:
if not is_otel_available():
raise ValueError(
"OpenTelemetry is not available. Unable to initialize "
"a tracer. Ensure OpenTelemetry packages are installed. "
f"Original error:\n{otel_import_error_traceback}")
f"Original error:\n{otel_import_error_traceback}"
)
trace_provider = TracerProvider()
span_exporter = get_span_exporter(otlp_traces_endpoint)
@@ -70,19 +76,19 @@ def get_span_exporter(endpoint):
protocol = os.environ.get(OTEL_EXPORTER_OTLP_TRACES_PROTOCOL, "grpc")
if protocol == "grpc":
from opentelemetry.exporter.otlp.proto.grpc.trace_exporter import (
OTLPSpanExporter)
OTLPSpanExporter,
)
elif protocol == "http/protobuf":
from opentelemetry.exporter.otlp.proto.http.trace_exporter import (
OTLPSpanExporter) # type: ignore
OTLPSpanExporter, # type: ignore
)
else:
raise ValueError(
f"Unsupported OTLP protocol '{protocol}' is configured")
raise ValueError(f"Unsupported OTLP protocol '{protocol}' is configured")
return OTLPSpanExporter(endpoint=endpoint)
def extract_trace_context(
headers: Optional[Mapping[str, str]]) -> Optional[Context]:
def extract_trace_context(headers: Optional[Mapping[str, str]]) -> Optional[Context]:
if is_otel_available():
headers = headers or {}
return TraceContextTextMapPropagator().extract(headers)
@@ -91,7 +97,6 @@ def extract_trace_context(
def extract_trace_headers(headers: Mapping[str, str]) -> Mapping[str, str]:
return {h: headers[h] for h in TRACE_HEADERS if h in headers}
@@ -113,17 +118,13 @@ class SpanAttributes:
GEN_AI_LATENCY_E2E = "gen_ai.latency.e2e"
GEN_AI_LATENCY_TIME_IN_SCHEDULER = "gen_ai.latency.time_in_scheduler"
# Time taken in the forward pass for this across all workers
GEN_AI_LATENCY_TIME_IN_MODEL_FORWARD = (
"gen_ai.latency.time_in_model_forward")
GEN_AI_LATENCY_TIME_IN_MODEL_FORWARD = "gen_ai.latency.time_in_model_forward"
# Time taken in the model execute function. This will include model
# forward, block/sync across workers, cpu-gpu sync time and sampling time.
GEN_AI_LATENCY_TIME_IN_MODEL_EXECUTE = (
"gen_ai.latency.time_in_model_execute")
GEN_AI_LATENCY_TIME_IN_MODEL_PREFILL = \
"gen_ai.latency.time_in_model_prefill"
GEN_AI_LATENCY_TIME_IN_MODEL_EXECUTE = "gen_ai.latency.time_in_model_execute"
GEN_AI_LATENCY_TIME_IN_MODEL_PREFILL = "gen_ai.latency.time_in_model_prefill"
GEN_AI_LATENCY_TIME_IN_MODEL_DECODE = "gen_ai.latency.time_in_model_decode"
GEN_AI_LATENCY_TIME_IN_MODEL_INFERENCE = \
"gen_ai.latency.time_in_model_inference"
GEN_AI_LATENCY_TIME_IN_MODEL_INFERENCE = "gen_ai.latency.time_in_model_inference"
def contains_trace_headers(headers: Mapping[str, str]) -> bool:
@@ -132,5 +133,4 @@ def contains_trace_headers(headers: Mapping[str, str]) -> bool:
@run_once
def log_tracing_disabled_warning() -> None:
logger.warning(
"Received a request with trace context but tracing is disabled")
logger.warning("Received a request with trace context but tracing is disabled")