feat(tracing): resource renaming

.github/workflows/system-tests.yml

@@ -47,7 +47,7 @@ jobs:
           persist-credentials: false
           repository: 'DataDog/system-tests'
           # Automatically managed, use scripts/update-system-tests-version to update
-          ref: '4bece37901d6241526eb788bdac6f2887251cdcd'
+          ref: 'da285213e99805433b99f8ea9d1e57f68e2b8e3d'
 
       - name: Download wheels to binaries directory
         uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
@@ -94,7 +94,7 @@ jobs:
           persist-credentials: false
           repository: 'DataDog/system-tests'
           # Automatically managed, use scripts/update-system-tests-version to update
-          ref: '4bece37901d6241526eb788bdac6f2887251cdcd'
+          ref: 'da285213e99805433b99f8ea9d1e57f68e2b8e3d'
 
       - name: Build runner
         uses: ./.github/actions/install_runner
@@ -283,7 +283,7 @@ jobs:
           persist-credentials: false
           repository: 'DataDog/system-tests'
           # Automatically managed, use scripts/update-system-tests-version to update
-          ref: '4bece37901d6241526eb788bdac6f2887251cdcd'
+          ref: 'da285213e99805433b99f8ea9d1e57f68e2b8e3d'
       - name: Download wheels to binaries directory
         uses: actions/download-artifact@d3f86a106a0bac45b974a628896c90dbdf5c8093 # v4.3.0
         with:

.gitlab-ci.yml

@@ -14,7 +14,7 @@ variables:
   DD_VPA_TEMPLATE: "vpa-template-cpu-p70-10percent-2x-oom-min-cap"
   # CI_DEBUG_SERVICES: "true"
   # Automatically managed, use scripts/update-system-tests-version to update
-  SYSTEM_TESTS_REF: "4bece37901d6241526eb788bdac6f2887251cdcd"
+  SYSTEM_TESTS_REF: "da285213e99805433b99f8ea9d1e57f68e2b8e3d"
 
 default:
   interruptible: true

benchmarks/bm/flask_utils.py

@@ -92,6 +92,7 @@ def setup(self):
                 "DD_APPSEC_ENABLED": str(self.appsec_enabled),
                 "DD_IAST_ENABLED": str(self.iast_enabled),
                 "DD_TELEMETRY_METRICS_ENABLED": str(self.telemetry_metrics_enabled),
+                "DD_TRACE_RESOURCE_RENAMING_ENABLED": str(self.resource_renaming_enabled),
             }
         )
 

benchmarks/django_simple/config.yaml

@@ -11,6 +11,7 @@ baseline: &baseline
   django_instrument_templates: true
   native_writer: false
   django_minimal: false
+  resource_renaming_enabled: false
 tracer: &tracer
   <<: *baseline
   tracer_enabled: true
@@ -56,3 +57,6 @@ tracer-native:
 tracer-minimal:
   <<: *tracer
   django_minimal: true
+resource-renaming:
+  <<: *tracer
+  resource_renaming_enabled: true

benchmarks/django_simple/scenario.py

@@ -18,6 +18,7 @@ class DjangoSimple(bm.Scenario):
     django_instrument_templates: bool
     native_writer: bool
     django_minimal: bool
+    resource_renaming_enabled: bool
 
     def run(self):
         os.environ["DJANGO_SETTINGS_MODULE"] = "app"
@@ -49,6 +50,8 @@ def run(self):
             os.environ.update({"DD_EXCEPTION_REPLAY_ENABLED": "1"})
         if self.native_writer:
             os.environ.update({"_DD_TRACE_WRITER_NATIVE": "1"})
+        if self.resource_renaming_enabled:
+            os.environ.update({"DD_TRACE_RESOURCE_RENAMING_ENABLED": "1"})
 
         # This will not work with gevent workers as the gevent hub has not been
         # initialized when this hook is called.

benchmarks/errortracking_flask_sqli/scenario.py

@@ -14,6 +14,7 @@ class ErrorTrackingFlaskSQLi(bm.Scenario, FlaskScenarioMixin):
     telemetry_metrics_enabled: bool
     errortracking_enabled: str
     native_writer: bool
+    resource_renaming_enabled: bool
 
     def run(self):
         app = self.create_app()

benchmarks/flask_simple/config.yaml

@@ -7,6 +7,7 @@ baseline: &baseline
   post_request: false
   telemetry_metrics_enabled: false
   native_writer: false
+  resource_renaming_enabled: false
 tracer: &tracer
   <<: *baseline
   tracer_enabled: true
@@ -32,7 +33,10 @@ appsec-telemetry:
 tracer-native:
   <<: *tracer
   native_writer: true
-# The scenarios below produce inconsistent results. We plan to 
+resource-renaming:
+  <<: *tracer
+  resource_renaming_enabled: true
+# The scenarios below produce inconsistent results. We plan to
 # disbable these scenarios until the root cause is identified:
 # tracer-and-profiler:
 #   <<: *baseline

benchmarks/flask_simple/scenario.py

@@ -15,6 +15,7 @@ class FlaskSimple(bm.Scenario, FlaskScenarioMixin):
     telemetry_metrics_enabled: bool
     errortracking_enabled: str
     native_writer: bool
+    resource_renaming_enabled: bool
 
     def run(self):
         app = self.create_app()

benchmarks/flask_sqli/scenario.py

@@ -14,6 +14,7 @@ class FlaskSQLi(bm.Scenario, FlaskScenarioMixin):
     telemetry_metrics_enabled: bool
     errortracking_enabled: str
     native_writer: bool
+    resource_renaming_enabled: bool
 
     def run(self):
         app = self.create_app()

ddtrace/_trace/processor/resource_renaming.py

@@ -0,0 +1,77 @@
+import re
+from typing import List
+from typing import Optional
+from urllib.parse import urlparse
+
+from ddtrace._trace.processor import SpanProcessor
+from ddtrace.ext import SpanTypes
+from ddtrace.ext import http
+from ddtrace.internal.logger import get_logger
+from ddtrace.settings._config import config
+
+
+log = get_logger(__name__)
+
+
+class ResourceRenamingProcessor(SpanProcessor):
+    def __init__(self):
+        self._INT_RE = re.compile(r"^[1-9][0-9]+$")
+        self._INT_ID_RE = re.compile(r"^(?=.*[0-9].*)[0-9._-]{3,}$")
+        self._HEX_RE = re.compile(r"^(?=.*[0-9].*)[A-Fa-f0-9]{6,}$")
+        self._HEX_ID_RE = re.compile(r"^(?=.*[0-9].*)[A-Fa-f0-9._-]{6,}$")
+        self._STR_RE = re.compile(r"^(.{20,}|.*[%&'()*+,:=@].*)$")
+
+    def _compute_simplified_endpoint_path_element(self, elem: str) -> str:
+        """Applies the parameter replacement rules to a single path element."""
+        if self._INT_RE.fullmatch(elem):
+            return "{param:int}"
+        if self._INT_ID_RE.fullmatch(elem):
+            return "{param:int_id}"
+        if self._HEX_RE.fullmatch(elem):
+            return "{param:hex}"
+        if self._HEX_ID_RE.fullmatch(elem):
+            return "{param:hex_id}"
+        if self._STR_RE.fullmatch(elem):
+            return "{param:str}"
+        return elem
+
+    def _compute_simplified_endpoint(self, url: Optional[str]) -> str:
+        """Extracts and simplifies the path from an HTTP URL."""
+        if not url:
+            return "/"
+
+        try:
+            parsed_url = urlparse(url)
+        except ValueError as e:
+            log.error("Failed to parse http.url tag when processing span for resource renaming: %s", e)
+            return "/"
+        path = parsed_url.path
+        if not path or path == "/":
+            return "/"
+
+        elements: List[str] = []
+        for part in path.split("/"):
+            if part:
+                elements.append(part)
+                if len(elements) >= 8:
+                    break
+
+        if not elements:
+            return "/"
+
+        elements = [self._compute_simplified_endpoint_path_element(elem) for elem in elements]
+        return "/" + "/".join(elements)
+
+    def on_span_start(self, span):
+        pass
+
+    def on_span_finish(self, span):
+        if not span._is_top_level or span.span_type not in (SpanTypes.WEB, SpanTypes.HTTP, SpanTypes.SERVERLESS):
+            return
+
+        route = span.get_tag(http.ROUTE)
+
+        if not route or config._trace_resource_renaming_always_simplified_endpoint:
+            url = span.get_tag(http.URL)
+            endpoint = self._compute_simplified_endpoint(url)
+            span.set_tag_str(http.ENDPOINT, endpoint)

ddtrace/_trace/tracer.py

@@ -20,6 +20,7 @@
 from ddtrace._trace.processor import SpanProcessor
 from ddtrace._trace.processor import TopLevelSpanProcessor
 from ddtrace._trace.processor import TraceProcessor
+from ddtrace._trace.processor.resource_renaming import ResourceRenamingProcessor
 from ddtrace._trace.provider import BaseContextProvider
 from ddtrace._trace.provider import DefaultContextProvider
 from ddtrace._trace.span import Span
@@ -74,6 +75,9 @@ def _default_span_processors_factory(
     span_processors: List[SpanProcessor] = []
     span_processors += [TopLevelSpanProcessor()]
 
+    if config._trace_resource_renaming_enabled:
+        span_processors.append(ResourceRenamingProcessor())
+
     # When using the NativeWriter stats are computed by the native code.
     if config._trace_compute_stats and not config._trace_writer_native:
         # Inline the import to avoid pulling in ddsketch or protobuf

ddtrace/ext/http.py

@@ -18,6 +18,7 @@
 CLIENT_IP = "http.client_ip"
 ROUTE = "http.route"
 REFERRER_HOSTNAME = "http.referrer_hostname"
+ENDPOINT = "http.endpoint"
 
 # HTTP headers
 REFERER_HEADER = "referer"

ddtrace/internal/processor/stats.py

@@ -48,6 +48,8 @@ def _is_measured(span: Span) -> bool:
     str,  # type
     int,  # http status code
     bool,  # synthetics request
+    str,  # http method
+    str,  # http endpoint
 ]
 
 
@@ -73,8 +75,10 @@ def _span_aggr_key(span: Span) -> SpanAggrKey:
     resource = span.resource or ""
     _type = span.span_type or ""
     status_code = span.get_tag("http.status_code") or 0
+    method = span.get_tag("http.method") or ""
+    endpoint = span.get_tag("http.endpoint") or span.get_tag("http.route") or ""
     synthetics = span.context.dd_origin == "synthetics"
-    return span.name, service, resource, _type, int(status_code), synthetics
+    return (span.name, service, resource, _type, int(status_code), synthetics, method, endpoint)
 
 
 class SpanStatsProcessorV06(PeriodicService, SpanProcessor):
@@ -157,12 +161,14 @@ def _serialize_buckets(self) -> List[Dict]:
             serialized_bucket_keys.append(bucket_time_ns)
 
             for aggr_key, stat_aggr in bucket.items():
-                name, service, resource, _type, http_status, synthetics = aggr_key
+                name, service, resource, _type, http_status, synthetics, http_method, http_endpoint = aggr_key
                 serialized_bucket = {
                     "Name": compat.ensure_text(name),
                     "Resource": compat.ensure_text(resource),
                     "Synthetics": synthetics,
                     "HTTPStatusCode": http_status,
+                    "HTTPMethod": http_method,
+                    "HTTPEndpoint": http_endpoint,
                     "Hits": stat_aggr.hits,
                     "TopLevelHits": stat_aggr.top_level_hits,
                     "Duration": stat_aggr.duration,

ddtrace/settings/_config.py

@@ -661,6 +661,14 @@ def __init__(self):
         self._inferred_proxy_services_enabled = _get_config("DD_TRACE_INFERRED_PROXY_SERVICES_ENABLED", False, asbool)
         self._trace_safe_instrumentation_enabled = _get_config("DD_TRACE_SAFE_INSTRUMENTATION_ENABLED", False, asbool)
 
+        # Resource renaming
+        self._trace_resource_renaming_enabled = _get_config(
+            "DD_TRACE_RESOURCE_RENAMING_ENABLED", default=False, modifier=asbool
+        )
+        self._trace_resource_renaming_always_simplified_endpoint = _get_config(
+            "DD_TRACE_RESOURCE_RENAMING_ALWAYS_SIMPLIFIED_ENDPOINT", default=False, modifier=asbool
+        )
+
     def __getattr__(self, name) -> Any:
         if name in self._config:
             return self._config[name].value()

releasenotes/notes/feat-resource-renaming-e5911d3975c220e3.yaml

@@ -0,0 +1,4 @@
+---
+features:
+  - |
+    tracing: Added support for resource renaming, a experimental feature that enables automatic renaming of the resource field of spans for web requests from the Datadog user interface when the resource name does not correctly represent an endpoint.

tests/telemetry/test_writer.py

@@ -519,6 +519,8 @@ def test_app_started_event_configuration_override(test_agent_session, run_python
         {"name": "DD_TRACE_PROPAGATION_STYLE_INJECT", "origin": "env_var", "value": "tracecontext"},
         {"name": "DD_TRACE_RATE_LIMIT", "origin": "env_var", "value": 50},
         {"name": "DD_TRACE_REPORT_HOSTNAME", "origin": "default", "value": False},
+        {"name": "DD_TRACE_RESOURCE_RENAMING_ALWAYS_SIMPLIFIED_ENDPOINT", "origin": "default", "value": False},
+        {"name": "DD_TRACE_RESOURCE_RENAMING_ENABLED", "origin": "default", "value": False},
         {"name": "DD_TRACE_SAFE_INSTRUMENTATION_ENABLED", "origin": "default", "value": False},
         {
             "name": "DD_TRACE_SAMPLING_RULES",