Use standard OTel tracing error handling

Author: aabmass

instrumentation-genai/opentelemetry-instrumentation-vertexai/src/opentelemetry/instrumentation/vertexai/patch.py

@@ -21,7 +21,6 @@
     GenerateContentParams,
     get_genai_request_attributes,
     get_span_name,
-    handle_span_exception,
 )
 from opentelemetry.trace import SpanKind, Tracer
 
@@ -85,33 +84,26 @@ def extract_params(
             name=span_name,
             kind=SpanKind.CLIENT,
             attributes=span_attributes,
-            end_on_exit=False,
-        ) as span:
+        ) as _span:
             # TODO: emit request events
             # if span.is_recording():
             #     for message in kwargs.get("messages", []):
             #         event_logger.emit(
             #             message_to_event(message, capture_content)
             #         )
 
-            try:
-                result = wrapped(*args, **kwargs)
-                # TODO: handle streaming
-                # if is_streaming(kwargs):
-                #     return StreamWrapper(
-                #         result, span, event_logger, capture_content
-                #     )
+            result = wrapped(*args, **kwargs)
+            # TODO: handle streaming
+            # if is_streaming(kwargs):
+            #     return StreamWrapper(
+            #         result, span, event_logger, capture_content
+            #     )
 
-                # TODO: add response attributes and events
-                # if span.is_recording():
-                #     _set_response_attributes(
-                #         span, result, event_logger, capture_content
-                #     )
-                span.end()
-                return result
-
-            except Exception as error:
-                handle_span_exception(span, error)
-                raise
+            # TODO: add response attributes and events
+            # if span.is_recording():
+            #     _set_response_attributes(
+            #         span, result, event_logger, capture_content
+            #     )
+            return result
 
     return traced_method

instrumentation-genai/opentelemetry-instrumentation-vertexai/src/opentelemetry/instrumentation/vertexai/utils.py

@@ -29,11 +29,6 @@
 from opentelemetry.semconv._incubating.attributes import (
     gen_ai_attributes as GenAIAttributes,
 )
-from opentelemetry.semconv.attributes import (
-    error_attributes as ErrorAttributes,
-)
-from opentelemetry.trace import Span
-from opentelemetry.trace.status import Status, StatusCode
 from opentelemetry.util.types import AttributeValue
 
 if TYPE_CHECKING:
@@ -146,12 +141,3 @@ def get_span_name(span_attributes: Mapping[str, AttributeValue]):
     name = span_attributes.get(GenAIAttributes.GEN_AI_OPERATION_NAME, "")
     model = span_attributes.get(GenAIAttributes.GEN_AI_REQUEST_MODEL, "")
     return f"{name} {model}"
-
-
-def handle_span_exception(span: Span, error: Exception):
-    span.set_status(Status(StatusCode.ERROR, str(error)))
-    if span.is_recording():
-        span.set_attribute(
-            ErrorAttributes.ERROR_TYPE, type(error).__qualname__
-        )
-    span.end()

instrumentation-genai/opentelemetry-instrumentation-vertexai/tests/cassettes/test_vertexai_generate_content.yaml

@@ -43,7 +43,7 @@ interactions:
                 ]
               },
               "finishReason": 1,
-              "avgLogprobs": -0.005634686664531105
+              "avgLogprobs": -0.0054909539850134595
             }
           ],
           "usageMetadata": {

instrumentation-genai/opentelemetry-instrumentation-vertexai/tests/cassettes/test_vertexai_generate_content_error.yaml

@@ -0,0 +1,64 @@
+interactions:
+- request:
+    body: |-
+      {
+        "contents": [
+          {
+            "role": "user",
+            "parts": [
+              {
+                "text": "Say this is a test"
+              }
+            ]
+          }
+        ],
+        "generationConfig": {
+          "temperature": 1000.0
+        }
+      }
+    headers:
+      Accept:
+      - '*/*'
+      Accept-Encoding:
+      - gzip, deflate
+      Connection:
+      - keep-alive
+      Content-Length:
+      - '196'
+      Content-Type:
+      - application/json
+      User-Agent:
+      - python-requests/2.32.3
+    method: POST
+    uri: https://us-central1-aiplatform.googleapis.com/v1/projects/fake-project/locations/us-central1/publishers/google/models/gemini-1.5-flash-002:generateContent?%24alt=json%3Benum-encoding%3Dint
+  response:
+    body:
+      string: |-
+        {
+          "error": {
+            "code": 400,
+            "message": "Unable to submit request because it has a temperature value of 1000 but the supported range is from 0 (inclusive) to 2.0001 (exclusive). Update the value and try again.",
+            "status": "INVALID_ARGUMENT",
+            "details": [
+              {
+                "@type": "type.googleapis.com/google.rpc.DebugInfo",
+                "detail": "[ORIGINAL ERROR] generic::invalid_argument: Unable to submit request because it has a temperature value of 1000 but the supported range is from 0 (inclusive) to 2.0001 (exclusive). Update the value and try again. [google.rpc.error_details_ext] { message: \"Unable to submit request because it has a temperature value of 1000 but the supported range is from 0 (inclusive) to 2.0001 (exclusive). Update the value and try again.\" }"
+              }
+            ]
+          }
+        }
+    headers:
+      Content-Type:
+      - application/json; charset=UTF-8
+      Transfer-Encoding:
+      - chunked
+      Vary:
+      - Origin
+      - X-Origin
+      - Referer
+      content-length:
+      - '809'
+    status:
+      code: 400
+      message: Bad Request
+version: 1

instrumentation-genai/opentelemetry-instrumentation-vertexai/tests/test_chat_completions.py

@@ -1,4 +1,5 @@
 import pytest
+from google.api_core.exceptions import BadRequest
 from vertexai.generative_models import (
     Content,
     GenerationConfig,
@@ -10,6 +11,7 @@
 from opentelemetry.sdk.trace.export.in_memory_span_exporter import (
     InMemorySpanExporter,
 )
+from opentelemetry.trace import StatusCode
 
 
 @pytest.mark.vcr
@@ -34,6 +36,42 @@ def test_vertexai_generate_content(
     }
 
 
+@pytest.mark.vcr
+def test_vertexai_generate_content_error(
+    span_exporter: InMemorySpanExporter,
+    instrument_with_content: VertexAIInstrumentor,
+):
+    model = GenerativeModel("gemini-1.5-flash-002")
+    try:
+        # Temperature out of range causes error
+        model.generate_content(
+            [
+                Content(
+                    role="user", parts=[Part.from_text("Say this is a test")]
+                )
+            ],
+            generation_config=GenerationConfig(temperature=1000),
+        )
+    except BadRequest:
+        pass
+
+    spans = span_exporter.get_finished_spans()
+    assert len(spans) == 1
+    assert spans[0].name == "chat gemini-1.5-flash-002"
+    assert dict(spans[0].attributes) == {
+        "gen_ai.operation.name": "chat",
+        "gen_ai.request.model": "gemini-1.5-flash-002",
+        "gen_ai.request.temperature": 1000.0,
+        "gen_ai.system": "vertex_ai",
+    }
+    # Sets error status
+    assert spans[0].status.status_code == StatusCode.ERROR
+
+    # Records exception event
+    assert len(spans[0].events) == 1
+    assert spans[0].events[0].name == "exception"
+
+
 @pytest.mark.vcr()
 def test_chat_completion_extra_client_level_params(
     span_exporter, instrument_no_content