Add ensure_ascii=False to json.dumps() calls in telemetry tracer (#37)

Author: moritalous

src/strands/telemetry/tracer.py

@@ -315,7 +315,7 @@ def start_model_invoke_span(
             "gen_ai.system": "strands-agents",
             "agent.name": agent_name,
             "gen_ai.agent.name": agent_name,
-            "gen_ai.prompt": json.dumps(messages, cls=JSONEncoder),
+            "gen_ai.prompt": serialize(messages),
         }
 
         if model_id:
@@ -338,7 +338,7 @@ def end_model_invoke_span(
             error: Optional exception if the model call failed.
         """
         attributes: Dict[str, AttributeValue] = {
-            "gen_ai.completion": json.dumps(message["content"], cls=JSONEncoder),
+            "gen_ai.completion": serialize(message["content"]),
             "gen_ai.usage.prompt_tokens": usage["inputTokens"],
             "gen_ai.usage.completion_tokens": usage["outputTokens"],
             "gen_ai.usage.total_tokens": usage["totalTokens"],
@@ -360,10 +360,10 @@ def start_tool_call_span(
             The created span, or None if tracing is not enabled.
         """
         attributes: Dict[str, AttributeValue] = {
-            "gen_ai.prompt": json.dumps(tool, cls=JSONEncoder),
+            "gen_ai.prompt": serialize(tool),
             "tool.name": tool["name"],
             "tool.id": tool["toolUseId"],
-            "tool.parameters": json.dumps(tool["input"], cls=JSONEncoder),
+            "tool.parameters": serialize(tool["input"]),
         }
 
         # Add additional kwargs as attributes
@@ -387,7 +387,7 @@ def end_tool_call_span(
             status = tool_result.get("status")
             status_str = str(status) if status is not None else ""
 
-            tool_result_content_json = json.dumps(tool_result.get("content"), cls=JSONEncoder)
+            tool_result_content_json = serialize(tool_result.get("content"))
             attributes.update(
                 {
                     "tool.result": tool_result_content_json,
@@ -420,7 +420,7 @@ def start_event_loop_cycle_span(
         parent_span = parent_span if parent_span else event_loop_kwargs.get("event_loop_parent_span")
 
         attributes: Dict[str, AttributeValue] = {
-            "gen_ai.prompt": json.dumps(messages, cls=JSONEncoder),
+            "gen_ai.prompt": serialize(messages),
             "event_loop.cycle_id": event_loop_cycle_id,
         }
 
@@ -449,11 +449,11 @@ def end_event_loop_cycle_span(
             error: Optional exception if the cycle failed.
         """
         attributes: Dict[str, AttributeValue] = {
-            "gen_ai.completion": json.dumps(message["content"], cls=JSONEncoder),
+            "gen_ai.completion": serialize(message["content"]),
         }
 
         if tool_result_message:
-            attributes["tool.result"] = json.dumps(tool_result_message["content"], cls=JSONEncoder)
+            attributes["tool.result"] = serialize(tool_result_message["content"])
 
         self._end_span(span, attributes, error)
 
@@ -490,7 +490,7 @@ def start_agent_span(
             attributes["gen_ai.request.model"] = model_id
 
         if tools:
-            tools_json = json.dumps(tools, cls=JSONEncoder)
+            tools_json = serialize(tools)
             attributes["agent.tools"] = tools_json
             attributes["gen_ai.agent.tools"] = tools_json
 
@@ -571,3 +571,15 @@ def get_tracer(
         )
 
     return _tracer_instance
+
+
+def serialize(obj: Any) -> str:
+    """Serialize an object to JSON with consistent settings.
+
+    Args:
+        obj: The object to serialize
+
+    Returns:
+        JSON string representation of the object
+    """
+    return json.dumps(obj, ensure_ascii=False, cls=JSONEncoder)

tests/strands/telemetry/test_tracer.py

@@ -6,7 +6,7 @@
 import pytest
 from opentelemetry.trace import StatusCode  # type: ignore
 
-from strands.telemetry.tracer import JSONEncoder, Tracer, get_tracer
+from strands.telemetry.tracer import JSONEncoder, Tracer, get_tracer, serialize
 from strands.types.streaming import Usage
 
 
@@ -635,3 +635,49 @@ def test_json_encoder_value_error():
     # Test just the value
     result = json.loads(encoder.encode(huge_number))
     assert result == "<replaced>"
+
+
+def test_serialize_non_ascii_characters():
+    """Test that non-ASCII characters are preserved in JSON serialization."""
+
+    # Test with Japanese text
+    japanese_text = "こんにちは世界"
+    result = serialize({"text": japanese_text})
+    assert japanese_text in result
+    assert "\\u" not in result
+
+    # Test with emoji
+    emoji_text = "Hello 🌍"
+    result = serialize({"text": emoji_text})
+    assert emoji_text in result
+    assert "\\u" not in result
+
+    # Test with Chinese characters
+    chinese_text = "你好，世界"
+    result = serialize({"text": chinese_text})
+    assert chinese_text in result
+    assert "\\u" not in result
+
+    # Test with mixed content
+    mixed_text = {"ja": "こんにちは", "emoji": "😊", "zh": "你好", "en": "hello"}
+    result = serialize(mixed_text)
+    assert "こんにちは" in result
+    assert "😊" in result
+    assert "你好" in result
+    assert "\\u" not in result
+
+
+def test_serialize_vs_json_dumps():
+    """Test that serialize behaves differently from default json.dumps for non-ASCII characters."""
+
+    # Test with Japanese text
+    japanese_text = "こんにちは世界"
+
+    # Default json.dumps should escape non-ASCII characters
+    default_result = json.dumps({"text": japanese_text})
+    assert "\\u" in default_result
+
+    # Our serialize function should preserve non-ASCII characters
+    custom_result = serialize({"text": japanese_text})
+    assert japanese_text in custom_result
+    assert "\\u" not in custom_result