add trace

2026-05-06 13:56:41 +08:00
parent 710fbc32be
commit f8bfdbd92a
8 changed files with 194 additions and 51 deletions
--- a/README.md
+++ b/README.md
@@ -3,16 +3,18 @@

 ```mermaid
 flowchart LR
-    SDK[Python SDK] -->|logs OTLP/HTTP| FB[Fluent Bit]
+    SDK[Python SDK] -->|logs/traces OTLP/HTTP| FB[Fluent Bit]
    SDK -->|metrics Prometheus remote_write| FB

    FB -->|logs| VL[VictoriaLogs]
+    FB -->|traces| VT[VictoriaTraces]

    FB -->|metrics remote_write| VM[VictoriaMetrics]

    VM -->|query| GF[Grafana]

    VL -->|query| GF
+    VT -->|Jaeger query| GF
 ```


@@ -30,9 +32,9 @@ flowchart TD

    BK[Backend] -->|trace| FB

-    FB -->|logs| VL[VictoriaLogs cluster]
-    FB -->|metrics remote_write| VM[VictoriaMetrics]
-    FB -->|trace| VT[VictoriaTraces]
+    FB -->|logs| VL[VictoriaLogs &#40cluster&#41]
+    FB -->|metrics remote_write| VM[VictoriaMetrics &#40cluster&#41]
+    FB -->|trace| VT[VictoriaTraces &#40cluster&#41]

    VM -->|query| GF[Grafana]
    VL -->|query| GF
--- a/demo-app/main.py
+++ b/demo-app/main.py
@@ -34,7 +34,7 @@ def main() -> None:
            except Exception as exc:
                print(f"ERROR path={path!r} type={exc.__class__.__name__} error={exc}")

-        # Demo app is short-looping, so flush logs and remote_write metrics quickly.
+        # Demo app is short-looping, so flush logs, traces, and remote_write metrics quickly.
        force_flush()

        if not loop:
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -64,6 +64,29 @@ services:
        source: victoriametrics
        target: /vmsingle
        volume: {}
+  victoriatraces:
+    command:
+      - -storageDataPath=/vtraces
+      - -loggerFormat=json
+    healthcheck:
+      test:
+        - CMD
+        - wget
+        - -qO-
+        - http://127.0.0.1:10428/health
+      timeout: 1s
+      interval: 1s
+      retries: 10
+    image: docker.io/victoriametrics/victoria-traces:latest
+    networks:
+      default: null
+    ports:
+      - 10428:10428
+    volumes:
+      - type: volume
+        source: victoriatraces
+        target: /vtraces
+        volume: {}
  demo-app:
    build:
      context: .
@@ -73,6 +96,7 @@ services:
      STARRYSDK_SERVICE_NAME: starry-python-sdk-demo
      STARRYSDK_PROMETHEUS_REMOTE_WRITE_ENDPOINT: http://fluentbit:8080/api/prom/push
      STARRYSDK_METRIC_PUSH_INTERVAL_SECONDS: "5"
+      STARRYSDK_TRACE_SAMPLE_RATE: "1.0"
      OTEL_EXPORTER_OTLP_ENDPOINT: http://fluentbit:4318
      DEMO_LOOP: "true"
      DEMO_INTERVAL_SECONDS: "5"
@@ -92,5 +116,7 @@ volumes:
    name: fluentbit-oltp_victorialogs-1
  victoriametrics:
    name: fluentbit-oltp_victoriametrics
+  victoriatraces:
+    name: fluentbit-oltp_victoriatraces
  grafana-data:
    name: fluentbit-oltp_grafana
--- a/fluent-bit.conf
+++ b/fluent-bit.conf
@@ -67,6 +67,18 @@
    Compress              gzip
    Workers               2

+# Python SDK 发来的 traces -> VictoriaTraces
+[OUTPUT]
+    Name                  opentelemetry
+    Match                 v1_traces
+    Host                  victoriatraces
+    Port                  10428
+
+    Traces_Uri            /insert/opentelemetry/v1/traces
+
+    Compress              gzip
+    Workers               2
+
 [OUTPUT]
    Name   stdout
    Match  *
--- a/starry-sdk/README.md
+++ b/starry-sdk/README.md
@@ -1,6 +1,6 @@
 # starry-client-sdk demo

-Demo SDK: request `https://blog.starryskymeow.top/{path}`, return response text, and raise `StarryNotFoundError` on HTTP 404. Telemetry is enabled by default. Metrics are pushed to Fluent Bit with Prometheus remote_write; exception logs are exported with OTLP/HTTP.
+Demo SDK: request `https://blog.starryskymeow.top/{path}`, return response text, and raise `StarryNotFoundError` on HTTP 404. Telemetry is enabled by default. Metrics are pushed to Fluent Bit with Prometheus remote_write; exception logs and traces are exported with OTLP/HTTP. In the demo compose stack, traces are forwarded from Fluent Bit to VictoriaTraces. Trace sampling defaults to 100%.

 Environment variables:

@@ -9,5 +9,7 @@ Environment variables:
 - `STARRYSDK_PROMETHEUS_REMOTE_WRITE_ENDPOINT`: default `http://host.docker.internal:8080/api/prom/push`.
 - `STARRYSDK_METRIC_PUSH_INTERVAL_SECONDS`: default `5`.
 - `STARRYSDK_METRIC_PUSH_TIMEOUT_SECONDS`: default `2`.
- `OTEL_EXPORTER_OTLP_ENDPOINT`: default `http://host.docker.internal:4318`; SDK appends `/v1/logs` for OTLP/HTTP logs.
+- `STARRYSDK_TRACE_SAMPLE_RATE`: default `1.0`; clamped to `0.0`-`1.0`.
+- `OTEL_EXPORTER_OTLP_ENDPOINT`: default `http://host.docker.internal:4318`; SDK appends `/v1/logs` and `/v1/traces` for OTLP/HTTP telemetry.
 - `OTEL_EXPORTER_OTLP_LOGS_ENDPOINT`: optional explicit logs endpoint.
+- `OTEL_EXPORTER_OTLP_TRACES_ENDPOINT`: optional explicit traces endpoint.
--- a/starry-sdk/pyproject.toml
+++ b/starry-sdk/pyproject.toml
@@ -5,7 +5,7 @@ build-backend = "setuptools.build_meta"
 [project]
 name = "starry-client-sdk"
 version = "0.1.0"
-description = "Demo Python SDK with default-on Prometheus metrics and OpenTelemetry exception logs."
+description = "Demo Python SDK with default-on Prometheus metrics, OpenTelemetry exception logs, and traces."
 readme = "README.md"
 requires-python = ">=3.9"
 dependencies = [
--- a/starry-sdk/starry_client_sdk/client.py
+++ b/starry-sdk/starry_client_sdk/client.py
@@ -4,6 +4,7 @@ import time
 from dataclasses import dataclass
 from typing import Any, Optional

+from opentelemetry.trace import SpanKind, Status, StatusCode
 import requests

 from .telemetry import SDK_NAME, configure_telemetry
@@ -89,10 +90,24 @@ class StarryClient:
            "error_type": "none",
        }

+        with self._telemetry.tracer.start_as_current_span(
+            "starry_client_sdk.get",
+            kind=SpanKind.CLIENT,
+            attributes={
+                "sdk.name": SDK_NAME,
+                "sdk.version": self.sdk_version,
+                "sdk.interface": context.sdk_interface,
+                "http.request.method": context.http_method,
+                "url.path": attrs["url_path"],
+            },
+            record_exception=False,
+            set_status_on_exception=False,
+        ) as span:
            start = time.perf_counter()
            try:
                response = requests.get(url, timeout=self.timeout_seconds)
                attrs["http_status_code"] = response.status_code
+                span.set_attribute("http.response.status_code", response.status_code)

                if response.status_code == 404:
                    attrs["outcome"] = "error"
@@ -101,10 +116,16 @@ class StarryClient:

                response.raise_for_status()
                attrs["outcome"] = "success"
+                span.set_attribute("sdk.outcome", attrs["outcome"])
+                span.set_status(Status(StatusCode.OK))
                return response.text
            except Exception as exc:
                attrs["outcome"] = "error"
                attrs["error_type"] = exc.__class__.__name__
+                span.set_attribute("sdk.outcome", attrs["outcome"])
+                span.set_attribute("error.type", attrs["error_type"])
+                span.record_exception(exc)
+                span.set_status(Status(StatusCode.ERROR, attrs["error_type"]))
                self._telemetry.logger.exception(
                    "Starry SDK request error",
                    extra={
@@ -120,6 +141,7 @@ class StarryClient:
                raise
            finally:
                duration_seconds = time.perf_counter() - start
+                span.set_attribute("sdk.duration_seconds", duration_seconds)
                self._telemetry.request_counter.add(1, attributes=attrs)
                self._telemetry.duration_histogram.record(duration_seconds, attributes=attrs)
                if attrs.get("outcome") == "error":
--- a/starry-sdk/starry_client_sdk/telemetry.py
+++ b/starry-sdk/starry_client_sdk/telemetry.py
@@ -34,6 +34,7 @@ _DURATION_BUCKETS = (0.005, 0.01, 0.025, 0.05, 0.075, 0.1, 0.25, 0.5, 0.75, 1.0,
 _LOCK = threading.Lock()
 _HANDLES: Optional["TelemetryHandles"] = None
 _LOGGER_PROVIDER: Any = None
+_TRACER_PROVIDER: Any = None
 _METRIC_PUSHER: Optional["_PrometheusRemoteWritePusher"] = None
 _WRITE_REQUEST_CLASS: Any = None

@@ -48,6 +49,28 @@ class _NoopHistogram:
        return None


+class _NoopSpan:
+    def __enter__(self) -> "_NoopSpan":
+        return self
+
+    def __exit__(self, exc_type: Any, exc: Any, traceback: Any) -> None:
+        return None
+
+    def set_attribute(self, key: str, value: Any) -> None:
+        return None
+
+    def record_exception(self, exception: Exception) -> None:
+        return None
+
+    def set_status(self, status: Any) -> None:
+        return None
+
+
+class _NoopTracer:
+    def start_as_current_span(self, name: str, **kwargs: Any) -> _NoopSpan:
+        return _NoopSpan()
+
+
 class _PrometheusCounter:
    def __init__(self, metric: Counter, static_labels: dict[str, str]) -> None:
        self._metric = metric
@@ -71,6 +94,7 @@ class TelemetryHandles:
    request_counter: Any
    error_counter: Any
    duration_histogram: Any
+    tracer: Any
    logger: logging.Logger
    enabled: bool

@@ -91,6 +115,10 @@ def _logs_endpoint() -> str:
    return os.getenv("OTEL_EXPORTER_OTLP_LOGS_ENDPOINT", f"{_base_endpoint()}/v1/logs")


+def _traces_endpoint() -> str:
+    return os.getenv("OTEL_EXPORTER_OTLP_TRACES_ENDPOINT", f"{_base_endpoint()}/v1/traces")
+
+
 def _remote_write_endpoint() -> str:
    return os.getenv("STARRYSDK_PROMETHEUS_REMOTE_WRITE_ENDPOINT", DEFAULT_PROMETHEUS_REMOTE_WRITE_ENDPOINT)

@@ -271,6 +299,7 @@ def _noop_handles() -> TelemetryHandles:
        request_counter=_NoopCounter(),
        error_counter=_NoopCounter(),
        duration_histogram=_NoopHistogram(),
+        tracer=_NoopTracer(),
        logger=logging.getLogger("starry_client_sdk"),
        enabled=False,
    )
@@ -352,12 +381,40 @@ def _configure_otel_logs(*, logger: logging.Logger, service_name: str, sdk_versi
        logger.addHandler(otel_handler)


+def _configure_otel_traces(*, service_name: str, sdk_version: str) -> Any:
+    global _TRACER_PROVIDER
+
+    from opentelemetry.exporter.otlp.proto.http.trace_exporter import OTLPSpanExporter
+    from opentelemetry.sdk.resources import Resource
+    from opentelemetry.sdk.trace import TracerProvider
+    from opentelemetry.sdk.trace.export import BatchSpanProcessor
+    from opentelemetry.sdk.trace.sampling import ParentBased, TraceIdRatioBased
+
+    sample_rate = max(0.0, min(1.0, _float_env("STARRYSDK_TRACE_SAMPLE_RATE", 1.0)))
+    resource = Resource.create(
+        {
+            "service.name": service_name,
+            "sdk.name": SDK_NAME,
+            "sdk.version": sdk_version,
+            "telemetry.source": "client-sdk",
+        }
+    )
+
+    _TRACER_PROVIDER = TracerProvider(
+        resource=resource,
+        sampler=ParentBased(root=TraceIdRatioBased(sample_rate)),
+    )
+    trace_exporter = OTLPSpanExporter(endpoint=_traces_endpoint())
+    _TRACER_PROVIDER.add_span_processor(BatchSpanProcessor(trace_exporter))
+    return _TRACER_PROVIDER.get_tracer(SDK_NAME, sdk_version)
+
+
 def configure_telemetry(*, service_name: Optional[str] = None, sdk_version: str = __version__) -> TelemetryHandles:
    """Configure default-on, non-blocking telemetry for this SDK.

-    Metrics are pushed to Fluent Bit with Prometheus remote_write. Exception logs are still exported with
-    OTLP/HTTP. Telemetry failures must never break business calls; if setup fails, the SDK falls back to
-    no-op meters.
+    Metrics are pushed to Fluent Bit with Prometheus remote_write. Exception logs and traces are exported
+    with OTLP/HTTP. Telemetry failures must never break business calls; if setup fails, the SDK falls back
+    to no-op handles.
    """

    global _HANDLES
@@ -377,7 +434,9 @@ def configure_telemetry(*, service_name: Optional[str] = None, sdk_version: str
        request_counter: Any = _NoopCounter()
        error_counter: Any = _NoopCounter()
        duration_histogram: Any = _NoopHistogram()
+        tracer: Any = _NoopTracer()
        metrics_enabled = False
+        traces_enabled = False

        try:
            request_counter, error_counter, duration_histogram = _configure_prometheus_remote_write(
@@ -396,12 +455,19 @@ def configure_telemetry(*, service_name: Optional[str] = None, sdk_version: str
        except Exception:
            logging.getLogger("starry_client_sdk.telemetry").debug("SDK OTLP log setup failed", exc_info=True)

+        try:
+            tracer = _configure_otel_traces(service_name=service_name, sdk_version=sdk_version)
+            traces_enabled = True
+        except Exception:
+            logging.getLogger("starry_client_sdk.telemetry").debug("SDK OTLP trace setup failed", exc_info=True)
+
        _HANDLES = TelemetryHandles(
            request_counter=request_counter,
            error_counter=error_counter,
            duration_histogram=duration_histogram,
+            tracer=tracer,
            logger=logger,
-            enabled=metrics_enabled,
+            enabled=metrics_enabled or traces_enabled,
        )
        atexit.register(shutdown_telemetry)
        return _HANDLES
@@ -414,11 +480,18 @@ def force_flush(timeout_millis: int = 5000) -> None:
        _METRIC_PUSHER.force_flush()

    if _LOGGER_PROVIDER is None:
-        return
+        pass
+    else:
        try:
            _LOGGER_PROVIDER.force_flush(timeout_millis=timeout_millis)
        except Exception:
-        logging.getLogger("starry_client_sdk.telemetry").debug("Telemetry force_flush failed", exc_info=True)
+            logging.getLogger("starry_client_sdk.telemetry").debug("Telemetry log force_flush failed", exc_info=True)
+
+    if _TRACER_PROVIDER is not None:
+        try:
+            _TRACER_PROVIDER.force_flush(timeout_millis=timeout_millis)
+        except Exception:
+            logging.getLogger("starry_client_sdk.telemetry").debug("Telemetry trace force_flush failed", exc_info=True)


 def shutdown_telemetry() -> None:
@@ -436,4 +509,10 @@ def shutdown_telemetry() -> None:
        try:
            _LOGGER_PROVIDER.shutdown()
        except Exception:
-            logging.getLogger("starry_client_sdk.telemetry").debug("Telemetry shutdown failed", exc_info=True)
+            logging.getLogger("starry_client_sdk.telemetry").debug("Telemetry log shutdown failed", exc_info=True)
+
+    if _TRACER_PROVIDER is not None:
+        try:
+            _TRACER_PROVIDER.shutdown()
+        except Exception:
+            logging.getLogger("starry_client_sdk.telemetry").debug("Telemetry trace shutdown failed", exc_info=True)