Add a resource detector for populating service.instance.id

CHANGELOG.md

@@ -46,6 +46,8 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
   ([#3991](https://github.com/open-telemetry/opentelemetry-python/pull/3991))
 - Add attributes field in  `MeterProvider.get_meter` and `InstrumentationScope`
   ([#4015](https://github.com/open-telemetry/opentelemetry-python/pull/4015))
+- Add `ServiceInstanceIdResourceDetector` that adds the 'service.instance.id' resource attribute
+  ([#4061](https://github.com/open-telemetry/opentelemetry-python/pull/4061))
 
 ## Version 1.25.0/0.46b0 (2024-05-30)
 

opentelemetry-sdk/pyproject.toml

@@ -67,6 +67,7 @@ console = "opentelemetry.sdk.trace.export:ConsoleSpanExporter"
 [project.entry-points.opentelemetry_resource_detector]
 otel = "opentelemetry.sdk.resources:OTELResourceDetector"
 process = "opentelemetry.sdk.resources:ProcessResourceDetector"
+serviceinstanceid = "opentelemetry.sdk.resources:ServiceInstanceIdResourceDetector"
 
 [project.urls]
 Homepage = "https://github.com/open-telemetry/opentelemetry-python/tree/main/opentelemetry-sdk"

opentelemetry-sdk/src/opentelemetry/sdk/resources/__init__.py

@@ -61,6 +61,7 @@
 import os
 import sys
 import typing
+import uuid
 from json import dumps
 from os import environ
 from urllib import parse
@@ -371,6 +372,17 @@ def detect(self) -> "Resource":
         return Resource(resource_info)
 
 
+class ServiceInstanceIdResourceDetector(ResourceDetector):
+    # pylint: disable=no-self-use
+    _instance_id_cache = {}
+
+    def detect(self) -> "Resource":
+        instance_id = self._instance_id_cache.setdefault(
+            os.getpid(), str(uuid.uuid4())
+        )
+        return Resource({SERVICE_INSTANCE_ID: instance_id})
+
+
 def get_aggregated_resources(
     detectors: typing.List["ResourceDetector"],
     initial_resource: typing.Optional[Resource] = None,

opentelemetry-sdk/tests/resources/test_resources.py

@@ -41,6 +41,7 @@
     PROCESS_RUNTIME_DESCRIPTION,
     PROCESS_RUNTIME_NAME,
     PROCESS_RUNTIME_VERSION,
+    SERVICE_INSTANCE_ID,
     SERVICE_NAME,
     TELEMETRY_SDK_LANGUAGE,
     TELEMETRY_SDK_NAME,
@@ -49,6 +50,7 @@
     ProcessResourceDetector,
     Resource,
     ResourceDetector,
+    ServiceInstanceIdResourceDetector,
     get_aggregated_resources,
 )
 
@@ -611,6 +613,27 @@ def test_process_detector(self):
             tuple(sys.argv),
         )
 
+    def test_service_instance_id_detector(self):
+        resource = get_aggregated_resources(
+            [ServiceInstanceIdResourceDetector()], Resource({})
+        )
+
+        self.assertIn(SERVICE_INSTANCE_ID, resource.attributes.keys())
+
+        # This throws if service instance id is not a valid UUID.
+        uuid.UUID(resource.attributes[SERVICE_INSTANCE_ID])
+
+        other_resource = get_aggregated_resources(
+            [ServiceInstanceIdResourceDetector()], Resource({})
+        )
+
+        # The instance id should be stable across invocations of the detector
+        # in the same process.
+        self.assertEqual(
+            resource.attributes[SERVICE_INSTANCE_ID],
+            other_resource.attributes[SERVICE_INSTANCE_ID],
+        )
+
     def test_resource_detector_entry_points_default(self):
         resource = Resource({}).create()
 
@@ -723,3 +746,15 @@ def test_resource_detector_entry_points_otel(self):
             )
             self.assertIn(PROCESS_RUNTIME_VERSION, resource.attributes.keys())
             self.assertEqual(resource.schema_url, "")
+        with patch.dict(
+            environ,
+            {
+                OTEL_EXPERIMENTAL_RESOURCE_DETECTORS: "serviceinstanceid",
+            },
+            clear=True,
+        ):
+            resource = Resource({}).create()
+            self.assertIn(SERVICE_INSTANCE_ID, resource.attributes.keys())
+
+            # This throws if service instance id is not a valid UUID.
+            uuid.UUID(resource.attributes[SERVICE_INSTANCE_ID])