feat: Modified internal span filtering logic according to new span-filtering mechanism

Signed-off-by: Cagri Yonca <cagri@ibm.com>

Author: CagriYonca

src/instana/instrumentation/urllib3.py

@@ -94,23 +94,7 @@ def urlopen_with_instana(
         tracer, parent_span, span_name = get_tracer_tuple()
 
         # If we're not tracing, just return; boto3 has it's own visibility
-        # Also, skip creating spans for internal Instana calls when
-        # 'com.instana' appears in either the full URL, the path argument,
-        # or the connection host.
-        request_url_or_path = (
-            kwargs.get("request_url")
-            or kwargs.get("url")
-            or (args[1] if len(args) >= 2 else "")
-            or ""
-        )
-        host = getattr(instance, "host", "") or ""
-
-        if (
-            not tracer
-            or span_name == "boto3"
-            or "com.instana" in request_url_or_path
-            or "com.instana" in host
-        ):
+        if not tracer or span_name == "boto3":
             return wrapped(*args, **kwargs)
 
         parent_context = parent_span.get_span_context() if parent_span else None

src/instana/options.py

@@ -134,6 +134,35 @@ def set_trace_configurations(self) -> None:
 
         self.set_disable_trace_configurations()
         self.set_stack_trace_configurations()
+        self._add_instana_agent_span_filter()
+
+    def _add_instana_agent_span_filter(self) -> None:
+        if "exclude" not in self.span_filters:
+            self.span_filters["exclude"] = []
+        self.span_filters["exclude"].extend(
+            [
+                {
+                    "name": "filter-internal-spans-by-url",
+                    "attributes": [
+                        {
+                            "key": "http.url",
+                            "values": ["com.instana"],
+                            "match_type": "contains",
+                        }
+                    ],
+                },
+                {
+                    "name": "filter-internal-spans-by-host",
+                    "attributes": [
+                        {
+                            "key": "http.host",
+                            "values": ["com.instana"],
+                            "match_type": "contains",
+                        }
+                    ],
+                },
+            ]
+        )
 
     def _apply_env_stack_trace_config(self) -> None:
         """Apply stack trace configuration from environment variables."""

src/instana/util/config.py

@@ -81,7 +81,7 @@ def parse_filtered_endpoints_string(params: Union[str, os.PathLike]) -> List[str
     return span_filters
 
 
-def parse_filtered_endpoints_dict(filter_dict: dict[str, Any]) -> dict[str, list[Any]]:
+def parse_filtered_endpoints_dict(filter_dict: Dict[str, Any]) -> Dict[str, List[Any]]:
     """
     Parses 'exclude' and 'include' blocks from the filter dict.
 
@@ -134,7 +134,7 @@ def parse_filtered_endpoints_dict(filter_dict: dict[str, Any]) -> dict[str, list
 
 def parse_filtered_endpoints(
     params: Union[Dict[str, Any], str],
-) -> Union[List[str], dict[str, list[Any]]]:
+) -> Union[List[str], Dict[str, List[Any]]]:
     """
     Parses input to prepare a list for ignored endpoints.
 
@@ -157,7 +157,7 @@ def parse_filtered_endpoints(
 
 def parse_filtered_endpoints_from_yaml(
     file_path: str,
-) -> Union[List[str], dict[str, list[Any]]]:
+) -> Union[List[str], Dict[str, List[Any]]]:
     """
     Parses configuration yaml file and prepares a list of ignored endpoints.
 
@@ -175,7 +175,7 @@ def parse_filtered_endpoints_from_yaml(
         span_filters = parse_filtered_endpoints(span_filters_dict)
         return span_filters
     else:
-        return []
+        return {}
 
 
 def parse_span_filter_env_vars() -> Dict[str, List[Any]]:

tests/clients/test_urllib3.py

@@ -1006,12 +1006,15 @@ def test_internal_span_creation_with_url_in_hostname(self) -> None:
 
         spans = self.recorder.queued_spans()
 
-        assert len(spans) == 1
+        assert len(spans) == 2
+
+        filtered_spans = agent.filter_spans(spans)
+        assert len(filtered_spans) == 1
 
-        test_span = spans[0]
+        test_span = filtered_spans[0]
         assert test_span.data["sdk"]["name"] == "test"
 
-        urllib3_spans = [span for span in spans if span.n == "urllib3"]
+        urllib3_spans = [span for span in filtered_spans if span.n == "urllib3"]
         assert len(urllib3_spans) == 0
 
     def test_internal_span_creation_with_url_in_path(self) -> None:
@@ -1024,11 +1027,13 @@ def test_internal_span_creation_with_url_in_path(self) -> None:
                 pass
 
         spans = self.recorder.queued_spans()
+        assert len(spans) == 2
 
-        assert len(spans) == 1
+        filtered_spans = agent.filter_spans(spans)
+        assert len(filtered_spans) == 1
 
-        test_span = spans[0]
+        test_span = filtered_spans[0]
         assert test_span.data["sdk"]["name"] == "test"
 
-        urllib3_spans = [span for span in spans if span.n == "urllib3"]
+        urllib3_spans = [span for span in filtered_spans if span.n == "urllib3"]
         assert len(urllib3_spans) == 0