fix: Modify span filtering configuration setting

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

Author: CagriYonca

src/instana/options.py

@@ -268,8 +268,11 @@ def set_span_filter_configurations(self) -> None:
         # The precedence is as follows:
         # environment variables > in-code configuration >
         # > agent config (configuration.yaml) > default value
-        if any(k.startswith("INSTANA_TRACING_FILTER_") for k in os.environ):
-            # Check for new span filtering env vars
+        if any(
+            k.startswith("INSTANA_TRACING_FILTER_") and os.environ[k]
+            for k in os.environ
+        ):
+            # Check for new span filtering env vars (only if at least one has a non-empty value)
             parsed_filter = parse_filter_rules_env_vars()
             if parsed_filter["exclude"] or parsed_filter["include"]:
                 self.span_filters = parsed_filter
@@ -393,8 +396,14 @@ def set_tracing(self, tracing: Dict[str, Any]) -> None:
         @param tracing: tracing configuration dictionary
         @return: None
         """
-        if "filter" in tracing and not self.span_filters:
-            self.span_filters = parse_filter_rules(tracing["filter"])
+        if "filter" in tracing and not self._has_high_priority_span_filter_source():
+            parsed = parse_filter_rules(tracing["filter"])
+            for policy in ("exclude", "include"):
+                rules = parsed.get(policy, [])
+                if rules:
+                    if policy not in self.span_filters:
+                        self.span_filters[policy] = []
+                    self.span_filters[policy].extend(rules)
 
         if "kafka" in tracing:
             if (
@@ -427,6 +436,21 @@ def set_tracing(self, tracing: Dict[str, Any]) -> None:
         # Handle stack trace configuration from agent config
         self.set_stack_trace_from_agent(tracing)
 
+    def _has_high_priority_span_filter_source(self) -> bool:
+        """Return True if a higher-priority span filter source (env var, YAML, or in-code config)
+        has already been configured, in which case the agent-provided filter should be ignored."""
+        return (
+            any(
+                k.startswith("INSTANA_TRACING_FILTER_") and os.environ[k]
+                for k in os.environ
+            )
+            or "INSTANA_CONFIG_PATH" in os.environ
+            or (
+                isinstance(config.get("tracing"), dict)
+                and "filter" in config["tracing"]
+            )
+        )
+
     def _should_apply_agent_global_config(self) -> bool:
         """Check if agent global config should be applied (lowest priority)."""
         has_env_vars = (

tests/test_options.py

@@ -689,8 +689,30 @@ def test_set_trace_configurations_by_agent_configuration(self) -> None:
         self.base_options = StandardOptions()
         self.base_options.set_tracing(test_tracing)
 
-        # set_tracing does not override span_filters when already set (has internal filters)
-        assert self.base_options.span_filters == {"exclude": INTERNAL_SPAN_FILTERS}
+        # Agent filter rules are appended after the internal filters (no high-priority source set).
+        agent_exclude = [
+            {
+                "name": "service1",
+                "suppression": True,
+                "attributes": [
+                    {"key": "service", "values": ["service1"], "match_type": "strict"}
+                ],
+            },
+            {
+                "name": "service2",
+                "suppression": True,
+                "attributes": [
+                    {
+                        "key": "method",
+                        "values": ["method1", "method2"],
+                        "match_type": "strict",
+                    }
+                ],
+            },
+        ]
+        assert self.base_options.span_filters == {
+            "exclude": INTERNAL_SPAN_FILTERS + agent_exclude
+        }
         assert self.base_options.kafka_trace_correlation
 
         # Check disabled_spans list
@@ -908,7 +930,28 @@ def test_set_tracing(
         }
         self.standart_options.set_tracing(test_tracing)
 
-        assert self.standart_options.span_filters == {"exclude": INTERNAL_SPAN_FILTERS}
+        # Agent filter rules are appended after the internal filters (no high-priority source set).
+        expected_exclude = INTERNAL_SPAN_FILTERS + [
+            {
+                "name": "service1",
+                "suppression": True,
+                "attributes": [
+                    {"key": "service", "values": ["service1"], "match_type": "strict"}
+                ],
+            },
+            {
+                "name": "service2",
+                "suppression": True,
+                "attributes": [
+                    {
+                        "key": "method",
+                        "values": ["method1", "method2"],
+                        "match_type": "strict",
+                    }
+                ],
+            },
+        ]
+        assert self.standart_options.span_filters == {"exclude": expected_exclude}
         assert not self.standart_options.kafka_trace_correlation
         assert (
             "Binary header format for Kafka is deprecated. Please use string header format."
@@ -972,7 +1015,30 @@ def test_set_from(self) -> None:
             self.standart_options.secrets_matcher == test_res_data["secrets"]["matcher"]
         )
         assert self.standart_options.secrets_list == test_res_data["secrets"]["list"]
-        assert self.standart_options.span_filters == {"exclude": INTERNAL_SPAN_FILTERS}
+        # Agent filter rules are appended after the internal filters.
+        agent_exclude = [
+            {
+                "name": "service1",
+                "suppression": True,
+                "attributes": [
+                    {"key": "service", "values": ["service1"], "match_type": "strict"}
+                ],
+            },
+            {
+                "name": "service2",
+                "suppression": True,
+                "attributes": [
+                    {
+                        "key": "method",
+                        "values": ["method1", "method2"],
+                        "match_type": "strict",
+                    }
+                ],
+            },
+        ]
+        assert self.standart_options.span_filters == {
+            "exclude": INTERNAL_SPAN_FILTERS + agent_exclude
+        }
 
         test_res_data2 = {
             "extraHeaders": {"header1": "sample-match", "header2": ["sample", "list"]},