Add some sem conv attributes to the stable vertex AI and gen AI instrumentations

instrumentation-genai/opentelemetry-instrumentation-google-genai/CHANGELOG.md

@@ -10,6 +10,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Ensure log event is written and completion hook is called even when model call results in exception. Put new
 log event (` gen_ai.client.inference.operation.details`) behind the flag `OTEL_SEMCONV_STABILITY_OPT_IN=gen_ai_latest_experimental`.
 Ensure same sem conv attributes are on the log and span. Fix an issue where the instrumentation would crash when a pydantic.BaseModel class was passed as the response schema ([#3905](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3905)).
+- Add the `GEN_AI_OUTPUT_TYPE` sem conv request attributes to events/spans generated in the stable instrumentation. This was added pre sem conv 1.36 so it should be in the stable instrumentation. Fix a bug in how system instructions were recorded in the `gen_ai.system.message` log event. It will now always be recorded as `{"content" : "text of system instructions"}`. See ([#4011](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/4011)).
 
 ## Version 0.4b0 (2025-10-16)
 

instrumentation-genai/opentelemetry-instrumentation-google-genai/src/opentelemetry/instrumentation/google_genai/generate_content.py

@@ -172,7 +172,6 @@ def _to_dict(value: object):
 
 def _create_request_attributes(
     config: Optional[GenerateContentConfigOrDict],
-    is_experimental_mode: bool,
     allow_list: AllowList,
 ) -> dict[str, Any]:
     if not config:
@@ -207,7 +206,7 @@ def _create_request_attributes(
         },
     )
     response_mime_type = config.get("response_mime_type")
-    if response_mime_type and is_experimental_mode:
+    if response_mime_type:
         if response_mime_type == "text/plain":
             attributes[gen_ai_attributes.GEN_AI_OUTPUT_TYPE] = "text"
         elif response_mime_type == "application/json":
@@ -505,31 +504,29 @@ def _maybe_log_completion_details(
     def _maybe_log_system_instruction(
         self, config: Optional[GenerateContentConfigOrDict] = None
     ):
-        system_instruction = None
-        if config is not None:
-            if isinstance(config, dict):
-                system_instruction = config.get("system_instruction")
-            else:
-                system_instruction = config.system_instruction
+        content_union = _config_to_system_instruction(config)
+        if not content_union:
+            return
+        content = transformers.t_contents(content_union)[0]
+        if not content.parts:
+            return
+        # System instruction is required to be text. An error will be returned by the API if it isn't.
+        system_instruction = " ".join(
+            part.text for part in content.parts if part.text
+        )
         if not system_instruction:
             return
-        attributes = {
-            gen_ai_attributes.GEN_AI_SYSTEM: self._genai_system,
-        }
-        # TODO: determine if "role" should be reported here or not. It is unclear
-        # since the caller does not supply a "role" and since this comes through
-        # a property named "system_instruction" which would seem to align with
-        # the default "role" that is allowed to be omitted by default.
-        #
-        # See also: "TODOS.md"
-        body = {}
-        if self._content_recording_enabled:
-            body["content"] = _to_dict(system_instruction)
-        else:
-            body["content"] = _CONTENT_ELIDED
         self._otel_wrapper.log_system_prompt(
-            attributes=attributes,
-            body=body,
+            attributes={
+                gen_ai_attributes.GEN_AI_SYSTEM: self._genai_system,
+            },
+            body={
+                "content": (
+                    system_instruction
+                    if self._content_recording_enabled
+                    else _CONTENT_ELIDED
+                )
+            },
         )
 
     def _maybe_log_user_prompt(
@@ -716,13 +713,8 @@ def instrumented_generate_content(
             completion_hook,
             generate_content_config_key_allowlist=generate_content_config_key_allowlist,
         )
-        is_experimental_mode = (
-            helper.sem_conv_opt_in_mode
-            == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
-        )
         request_attributes = _create_request_attributes(
             config,
-            is_experimental_mode,
             helper._generate_content_config_key_allowlist,
         )
         with helper.start_span_as_current_span(
@@ -739,7 +731,10 @@ def instrumented_generate_content(
                     config=helper.wrapped_config(config),
                     **kwargs,
                 )
-                if is_experimental_mode:
+                if (
+                    helper.sem_conv_opt_in_mode
+                    == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
+                ):
                     helper._update_response(response)
                     if response.candidates:
                         candidates += response.candidates
@@ -791,13 +786,8 @@ def instrumented_generate_content_stream(
             completion_hook,
             generate_content_config_key_allowlist=generate_content_config_key_allowlist,
         )
-        is_experimental_mode = (
-            helper.sem_conv_opt_in_mode
-            == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
-        )
         request_attributes = _create_request_attributes(
             config,
-            is_experimental_mode,
             helper._generate_content_config_key_allowlist,
         )
         with helper.start_span_as_current_span(
@@ -814,7 +804,10 @@ def instrumented_generate_content_stream(
                     config=helper.wrapped_config(config),
                     **kwargs,
                 ):
-                    if is_experimental_mode:
+                    if (
+                        helper.sem_conv_opt_in_mode
+                        == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
+                    ):
                         helper._update_response(response)
                         if response.candidates:
                             candidates += response.candidates
@@ -865,13 +858,8 @@ async def instrumented_generate_content(
             completion_hook,
             generate_content_config_key_allowlist=generate_content_config_key_allowlist,
         )
-        is_experimental_mode = (
-            helper.sem_conv_opt_in_mode
-            == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
-        )
         request_attributes = _create_request_attributes(
             config,
-            is_experimental_mode,
             helper._generate_content_config_key_allowlist,
         )
         candidates: list[Candidate] = []
@@ -889,7 +877,10 @@ async def instrumented_generate_content(
                     config=helper.wrapped_config(config),
                     **kwargs,
                 )
-                if is_experimental_mode:
+                if (
+                    helper.sem_conv_opt_in_mode
+                    == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
+                ):
                     helper._update_response(response)
                     if response.candidates:
                         candidates += response.candidates
@@ -940,13 +931,8 @@ async def instrumented_generate_content_stream(
             completion_hook,
             generate_content_config_key_allowlist=generate_content_config_key_allowlist,
         )
-        is_experimental_mode = (
-            helper.sem_conv_opt_in_mode
-            == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
-        )
         request_attributes = _create_request_attributes(
             config,
-            is_experimental_mode,
             helper._generate_content_config_key_allowlist,
         )
         with helper.start_span_as_current_span(
@@ -955,7 +941,10 @@ async def instrumented_generate_content_stream(
             end_on_exit=False,
         ) as span:
             span.set_attributes(request_attributes)
-            if not is_experimental_mode:
+            if (
+                not helper.sem_conv_opt_in_mode
+                == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
+            ):
                 helper.process_request(contents, config, span)
             try:
                 response_async_generator = await wrapped_func(
@@ -986,7 +975,10 @@ async def _response_async_generator_wrapper():
                 with trace.use_span(span, end_on_exit=True):
                     try:
                         async for response in response_async_generator:
-                            if is_experimental_mode:
+                            if (
+                                helper.sem_conv_opt_in_mode
+                                == _StabilityMode.GEN_AI_LATEST_EXPERIMENTAL
+                            ):
                                 helper._update_response(response)
                                 if response.candidates:
                                     candidates += response.candidates

instrumentation-genai/opentelemetry-instrumentation-google-genai/tests/generate_content/nonstreaming_base.py

@@ -17,7 +17,7 @@
 from unittest.mock import patch
 
 import pytest
-from google.genai.types import GenerateContentConfig
+from google.genai.types import GenerateContentConfig, Part
 from pydantic import BaseModel, Field
 
 from opentelemetry.instrumentation._semconv import (
@@ -180,6 +180,57 @@ def test_records_system_prompt_as_log(self):
         self.assertEqual(event_record.attributes["gen_ai.system"], "gemini")
         self.assertEqual(event_record.body["content"], "foo")
 
+    @patch.dict(
+        "os.environ",
+        {"OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT": "true"},
+    )
+    def test_system_prompt_passed_as_list_of_text(self):
+        config = GenerateContentConfig(
+            system_instruction=["help", "me please."]
+        )
+        self.configure_valid_response()
+        self.generate_content(
+            model="gemini-2.0-flash", contents="Some input", config=config
+        )
+        self.otel.assert_has_event_named("gen_ai.system.message")
+        event_record = self.otel.get_event_named("gen_ai.system.message")
+        self.assertEqual(event_record.body["content"], "help me please.")
+
+    @patch.dict(
+        "os.environ",
+        {"OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT": "true"},
+    )
+    def test_system_prompt_passed_as_list_of_text_parts(self):
+        config = GenerateContentConfig(
+            system_instruction=[
+                Part.from_text(text="help"),
+                Part.from_text(text="me please."),
+            ]
+        )
+        self.configure_valid_response()
+        self.generate_content(
+            model="gemini-2.0-flash", contents="Some input", config=config
+        )
+        self.otel.assert_has_event_named("gen_ai.system.message")
+        event_record = self.otel.get_event_named("gen_ai.system.message")
+        self.assertEqual(event_record.body["content"], "help me please.")
+
+    @patch.dict(
+        "os.environ",
+        {"OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT": "true"},
+    )
+    def test_system_prompt_passed_is_invalid(self):
+        config = GenerateContentConfig(
+            system_instruction=[
+                Part.from_uri(file_uri="test.jpg"),
+            ]
+        )
+        self.configure_valid_response()
+        self.generate_content(
+            model="gemini-2.0-flash", contents="Some input", config=config
+        )
+        self.otel.assert_does_not_have_event_named("gen_ai.system.message")
+
     @patch.dict(
         "os.environ",
         {"OTEL_INSTRUMENTATION_GENAI_CAPTURE_MESSAGE_CONTENT": "false"},

instrumentation-genai/opentelemetry-instrumentation-vertexai/CHANGELOG.md

@@ -8,6 +8,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 ## Unreleased
 - Fix overwritten log attributes in vertexai instrumentation
   ([#3925](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/3925))
+- Add the `GEN_AI_OUTPUT_TYPE`, `GEN_AI_REQUEST_CHOICE_COUNT` and `GEN_AI_REQUEST_SEED` sem conv request attributes to events/spans generated in the stable instrumentation. This was added pre sem conv 1.36 so it should be in the stable instrumentation. See ([#4011](https://github.com/open-telemetry/opentelemetry-python-contrib/pull/4011)).
 
 ## Version 2.1b0 (2025-10-16)
 

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

@@ -167,24 +167,23 @@ def get_genai_request_attributes(  # pylint: disable=too-many-branches
         attributes[GenAIAttributes.GEN_AI_REQUEST_STOP_SEQUENCES] = (
             generation_config.stop_sequences
         )
-    if use_latest_semconvs:
-        if "seed" in generation_config:
-            attributes[GenAIAttributes.GEN_AI_REQUEST_SEED] = (
-                generation_config.seed
-            )
-        if "candidate_count" in generation_config:
-            attributes[GenAIAttributes.GEN_AI_REQUEST_CHOICE_COUNT] = (
-                generation_config.candidate_count
+    if "seed" in generation_config:
+        attributes[GenAIAttributes.GEN_AI_REQUEST_SEED] = (
+            generation_config.seed
+        )
+    if "candidate_count" in generation_config:
+        attributes[GenAIAttributes.GEN_AI_REQUEST_CHOICE_COUNT] = (
+            generation_config.candidate_count
+        )
+    if "response_mime_type" in generation_config:
+        if generation_config.response_mime_type == "text/plain":
+            attributes[GenAIAttributes.GEN_AI_OUTPUT_TYPE] = "text"
+        elif generation_config.response_mime_type == "application/json":
+            attributes[GenAIAttributes.GEN_AI_OUTPUT_TYPE] = "json"
+        else:
+            attributes[GenAIAttributes.GEN_AI_OUTPUT_TYPE] = (
+                generation_config.response_mime_type
             )
-        if "response_mime_type" in generation_config:
-            if generation_config.response_mime_type == "text/plain":
-                attributes[GenAIAttributes.GEN_AI_OUTPUT_TYPE] = "text"
-            elif generation_config.response_mime_type == "application/json":
-                attributes[GenAIAttributes.GEN_AI_OUTPUT_TYPE] = "json"
-            else:
-                attributes[GenAIAttributes.GEN_AI_OUTPUT_TYPE] = (
-                    generation_config.response_mime_type
-                )
 
     return attributes
 

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

@@ -314,6 +314,7 @@ def test_generate_content_extra_params(
         "gen_ai.operation.name": "chat",
         "gen_ai.request.frequency_penalty": 1.0,
         "gen_ai.request.max_tokens": 5,
+        "gen_ai.request.seed": 12345,
         "gen_ai.request.model": "gemini-2.5-pro",
         "gen_ai.request.presence_penalty": -1.5,
         "gen_ai.request.stop_sequences": ("\n\n\n",),