ref: Replace set_data_normalized() with Span.set_data() for system instructions (#5374)

Author: alexander-alderman-webb

sentry_sdk/integrations/anthropic.py

@@ -1,4 +1,5 @@
 import sys
+import json
 from collections.abc import Iterable
 from functools import wraps
 from typing import TYPE_CHECKING
@@ -217,11 +218,9 @@ def _set_input_data(
         if isinstance(system_instructions, str) or isinstance(
             system_instructions, Iterable
         ):
-            set_data_normalized(
-                span,
+            span.set_data(
                 SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-                _transform_system_instructions(system_instructions),
-                unpack=False,
+                json.dumps(_transform_system_instructions(system_instructions)),
             )
 
         normalized_messages = []

sentry_sdk/integrations/google_genai/utils.py

@@ -1,4 +1,5 @@
 import copy
+import json
 import inspect
 from functools import wraps
 from .consts import ORIGIN, TOOL_ATTRIBUTES_MAP, GEN_AI_SYSTEM
@@ -808,11 +809,9 @@ def set_span_data_for_request(
             system_instructions = config.get("system_instruction")
 
         if system_instructions is not None:
-            set_data_normalized(
-                span,
+            span.set_data(
                 SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-                _transform_system_instructions(system_instructions),
-                unpack=False,
+                json.dumps(_transform_system_instructions(system_instructions)),
             )
 
         # Extract messages from contents

sentry_sdk/integrations/langchain.py

@@ -1,6 +1,7 @@
 import contextvars
 import itertools
 import sys
+import json
 import warnings
 from collections import OrderedDict
 from functools import wraps
@@ -467,11 +468,9 @@ def on_chat_model_start(
             if should_send_default_pii() and self.include_prompts:
                 system_instructions = _get_system_instructions(messages)
                 if len(system_instructions) > 0:
-                    set_data_normalized(
-                        span,
+                    span.set_data(
                         SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-                        _transform_system_instructions(system_instructions),
-                        unpack=False,
+                        json.dumps(_transform_system_instructions(system_instructions)),
                     )
 
                 normalized_messages = []

sentry_sdk/integrations/openai.py

@@ -1,4 +1,5 @@
 import sys
+import json
 import time
 from functools import wraps
 from collections.abc import Iterable
@@ -273,16 +274,16 @@ def _set_responses_api_input_data(
         and explicit_instructions is not None
         and _is_given(explicit_instructions)
     ):
-        set_data_normalized(
-            span,
+        span.set_data(
             SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-            [
-                {
-                    "type": "text",
-                    "content": explicit_instructions,
-                }
-            ],
-            unpack=False,
+            json.dumps(
+                [
+                    {
+                        "type": "text",
+                        "content": explicit_instructions,
+                    }
+                ]
+            ),
         )
 
         set_data_normalized(span, SPANDATA.GEN_AI_OPERATION_NAME, "responses")
@@ -309,11 +310,9 @@ def _set_responses_api_input_data(
     instructions_text_parts += _transform_system_instructions(system_instructions)
 
     if len(instructions_text_parts) > 0:
-        set_data_normalized(
-            span,
+        span.set_data(
             SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-            instructions_text_parts,
-            unpack=False,
+            json.dumps(instructions_text_parts),
         )
 
     if isinstance(messages, str):
@@ -365,11 +364,9 @@ def _set_completions_api_input_data(
 
     system_instructions = _get_system_instructions_completions(messages)
     if len(system_instructions) > 0:
-        set_data_normalized(
-            span,
+        span.set_data(
             SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-            _transform_system_instructions(system_instructions),
-            unpack=False,
+            json.dumps(_transform_system_instructions(system_instructions)),
         )
 
     if isinstance(messages, str):

sentry_sdk/integrations/openai_agents/utils.py

@@ -1,3 +1,5 @@
+import json
+
 import sentry_sdk
 from sentry_sdk.ai.utils import (
     GEN_AI_ALLOWED_MESSAGE_ROLES,
@@ -146,11 +148,9 @@ def _set_input_data(
     instructions_text_parts += _transform_system_instructions(system_instructions)
 
     if len(instructions_text_parts) > 0:
-        set_data_normalized(
-            span,
+        span.set_data(
             SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-            instructions_text_parts,
-            unpack=False,
+            json.dumps(instructions_text_parts),
         )
 
     non_system_messages = [

sentry_sdk/integrations/pydantic_ai/spans/ai_client.py

@@ -1,3 +1,5 @@
+import json
+
 import sentry_sdk
 from sentry_sdk._types import BLOB_DATA_SUBSTITUTE
 from sentry_sdk.ai.utils import (
@@ -101,13 +103,13 @@ def _set_input_messages(span: "sentry_sdk.tracing.Span", messages: "Any") -> Non
 
     permanent_instructions, current_instructions = _get_system_instructions(messages)
     if len(permanent_instructions) > 0 or len(current_instructions) > 0:
-        set_data_normalized(
-            span,
+        span.set_data(
             SPANDATA.GEN_AI_SYSTEM_INSTRUCTIONS,
-            _transform_system_instructions(
-                permanent_instructions, current_instructions
+            json.dumps(
+                _transform_system_instructions(
+                    permanent_instructions, current_instructions
+                )
             ),
-            unpack=False,
         )
 
     try: