Add back span status (#5147)

### Description
    
For some reason since Performance was implemented ages ago, we were not
sending the `span.status` field at all.
Add it back since we are moving to a Span-First world and Agent
Monitoring also expects span statuses to be correct for some product
features.
    
Some other notes regarding the status mess:
* the transaction status is added to the `trace_context` - this is fine
* the span status is also added as a tag, this should be removed in a
major since it's redundant
* `set_http_status` further adds another tag called `http.status_code`,
remove this in the major too

#### Issues

* resolves: #5059
* resolves: PY-1960

Author: sl0thentr0py

sentry_sdk/tracing.py

@@ -666,7 +666,7 @@ def set_http_status(self, http_status):
         # type: (int) -> None
         self.set_tag(
             "http.status_code", str(http_status)
-        )  # we keep this for backwards compatibility
+        )  # TODO-neel remove in major, we keep this for backwards compatibility
         self.set_data(SPANDATA.HTTP_STATUS_CODE, http_status)
         self.set_status(get_span_status_from_http_code(http_status))
 
@@ -729,6 +729,8 @@ def to_json(self):
         }  # type: Dict[str, Any]
 
         if self.status:
+            rv["status"] = self.status
+            # TODO-neel remove redundant tag in major
             self._tags["status"] = self.status
 
         if len(self._measurements) > 0:

tests/integrations/anthropic/test_anthropic.py

@@ -722,6 +722,7 @@ def test_span_status_error(sentry_init, capture_events):
 
     (error, transaction) = events
     assert error["level"] == "error"
+    assert transaction["spans"][0]["status"] == "internal_error"
     assert transaction["spans"][0]["tags"]["status"] == "internal_error"
     assert transaction["contexts"]["trace"]["status"] == "internal_error"
 
@@ -745,6 +746,7 @@ async def test_span_status_error_async(sentry_init, capture_events):
 
     (error, transaction) = events
     assert error["level"] == "error"
+    assert transaction["spans"][0]["status"] == "internal_error"
     assert transaction["spans"][0]["tags"]["status"] == "internal_error"
     assert transaction["contexts"]["trace"]["status"] == "internal_error"
 

tests/integrations/cohere/test_cohere.py

@@ -181,6 +181,7 @@ def test_span_status_error(sentry_init, capture_events):
 
     (error, transaction) = events
     assert error["level"] == "error"
+    assert transaction["spans"][0]["status"] == "internal_error"
     assert transaction["spans"][0]["tags"]["status"] == "internal_error"
     assert transaction["contexts"]["trace"]["status"] == "internal_error"
 

tests/integrations/huggingface_hub/test_huggingface_hub.py

@@ -792,6 +792,7 @@ def test_chat_completion_api_error(
     assert span["op"] == "gen_ai.chat"
     assert span["description"] == "chat test-model"
     assert span["origin"] == "auto.ai.huggingface_hub"
+    assert span["status"] == "internal_error"
     assert span.get("tags", {}).get("status") == "internal_error"
 
     assert (
@@ -835,6 +836,7 @@ def test_span_status_error(sentry_init, capture_events, mock_hf_api_with_errors)
             assert sp["op"] == "http.client"
 
     assert span is not None
+    assert span["status"] == "internal_error"
     assert span["tags"]["status"] == "internal_error"
 
     assert transaction["contexts"]["trace"]["status"] == "internal_error"

tests/integrations/langchain/test_langchain.py

@@ -347,6 +347,7 @@ def test_span_status_error(sentry_init, capture_events):
 
     (error, transaction) = events
     assert error["level"] == "error"
+    assert transaction["spans"][0]["status"] == "internal_error"
     assert transaction["spans"][0]["tags"]["status"] == "internal_error"
     assert transaction["contexts"]["trace"]["status"] == "internal_error"
 

tests/integrations/langgraph/test_langgraph.py

@@ -403,6 +403,7 @@ def original_invoke(self, *args, **kwargs):
     assert len(invoke_spans) == 1
 
     invoke_span = invoke_spans[0]
+    assert invoke_span.get("status") == "internal_error"
     assert invoke_span.get("tags", {}).get("status") == "internal_error"
 
 
@@ -436,6 +437,7 @@ async def run_error_test():
     assert len(invoke_spans) == 1
 
     invoke_span = invoke_spans[0]
+    assert invoke_span.get("status") == "internal_error"
     assert invoke_span.get("tags", {}).get("status") == "internal_error"
 
 

tests/integrations/mcp/test_mcp.py

@@ -300,6 +300,7 @@ def failing_tool(tool_name, arguments):
 
     # Error flag should be set for tools
     assert span["data"][SPANDATA.MCP_TOOL_RESULT_IS_ERROR] is True
+    assert span["status"] == "internal_error"
     assert span["tags"]["status"] == "internal_error"
 
 

tests/integrations/openai/test_openai.py

@@ -440,6 +440,7 @@ def test_span_status_error(sentry_init, capture_events):
 
     (error, transaction) = events
     assert error["level"] == "error"
+    assert transaction["spans"][0]["status"] == "internal_error"
     assert transaction["spans"][0]["tags"]["status"] == "internal_error"
     assert transaction["contexts"]["trace"]["status"] == "internal_error"
 

tests/integrations/openai_agents/test_openai_agents.py

@@ -645,6 +645,7 @@ async def test_error_handling(sentry_init, capture_events, test_agent):
 
     assert ai_client_span["description"] == "chat gpt-4"
     assert ai_client_span["origin"] == "auto.ai.openai_agents"
+    assert ai_client_span["status"] == "internal_error"
     assert ai_client_span["tags"]["status"] == "internal_error"
 
 
@@ -685,6 +686,7 @@ async def test_error_captures_input_data(sentry_init, capture_events, test_agent
     ai_client_span = [s for s in spans if s["op"] == "gen_ai.chat"][0]
 
     assert ai_client_span["description"] == "chat gpt-4"
+    assert ai_client_span["status"] == "internal_error"
     assert ai_client_span["tags"]["status"] == "internal_error"
 
     assert "gen_ai.request.messages" in ai_client_span["data"]
@@ -724,6 +726,7 @@ async def test_span_status_error(sentry_init, capture_events, test_agent):
 
     (error, transaction) = events
     assert error["level"] == "error"
+    assert transaction["spans"][0]["status"] == "internal_error"
     assert transaction["spans"][0]["tags"]["status"] == "internal_error"
     assert transaction["contexts"]["trace"]["status"] == "internal_error"
 
@@ -827,6 +830,7 @@ async def test_mcp_tool_execution_spans(sentry_init, capture_events, test_agent)
     )
 
     # Verify no error status since error was None
+    assert mcp_tool_span.get("status") != "internal_error"
     assert mcp_tool_span.get("tags", {}).get("status") != "internal_error"
 
 
@@ -927,6 +931,7 @@ async def test_mcp_tool_execution_with_error(sentry_init, capture_events, test_a
     assert mcp_tool_span["data"]["gen_ai.tool.output"] is None
 
     # Verify error status was set
+    assert mcp_tool_span["status"] == "internal_error"
     assert mcp_tool_span["tags"]["status"] == "internal_error"
 
 
@@ -1218,4 +1223,5 @@ def failing_tool(message: str) -> str:
 
     # Verify error status was set (this is the key test for our patch)
     # The span should be marked as error because the tool execution failed
+    assert execute_tool_span["status"] == "internal_error"
     assert execute_tool_span["tags"]["status"] == "internal_error"

tests/integrations/pymongo/test_pymongo.py

@@ -99,8 +99,11 @@ def test_transactions(sentry_init, capture_events, mongo_server, with_pii):
             and "4" not in insert_fail["description"]
         )
 
+    assert find["status"] == "ok"
     assert find["tags"]["status"] == "ok"
+    assert insert_success["status"] == "ok"
     assert insert_success["tags"]["status"] == "ok"
+    assert insert_fail["status"] == "internal_error"
     assert insert_fail["tags"]["status"] == "internal_error"