feat(profiling): Drop deprecated profiling sdks data (#91804)

This will start dropping profiling data from deprecated sdks when the
flag is enabled.

Author: Zylphrex

src/sentry/api/endpoints/organization_sdk_deprecations.py

@@ -2,19 +2,18 @@
 from typing import DefaultDict, TypedDict
 
 import sentry_sdk
-from packaging.version import InvalidVersion, Version
+from packaging.version import InvalidVersion
 from packaging.version import parse as parse_version
 from rest_framework import serializers
 from rest_framework.request import Request
 from rest_framework.response import Response
 
-from sentry import options
 from sentry.api.api_owners import ApiOwner
 from sentry.api.api_publish_status import ApiPublishStatus
 from sentry.api.base import region_silo_endpoint
 from sentry.api.bases.organization import OrganizationEndpoint
 from sentry.models.project import Project
-from sentry.models.projectsdk import EventType, ProjectSDK
+from sentry.models.projectsdk import EventType, ProjectSDK, get_minimum_sdk_version
 
 
 class SDKDeprecationsSerializer(serializers.Serializer):
@@ -93,40 +92,18 @@ def get_event_types(raw_event_type: str) -> list[EventType]:
     raise ValueError(f"Unknown event type: {raw_event_type}")
 
 
-MINIMUM_SDK_VERSION_OPTIONS: dict[tuple[int, str], str] = {
-    (EventType.PROFILE_CHUNK.value, "sentry.cocoa"): "sdk-deprecation.profile-chunk.cocoa",
-    (EventType.PROFILE_CHUNK.value, "sentry.python"): "sdk-deprecation.profile-chunk.python",
-}
-
-
-def get_minimum_sdk_version(project_sdk: ProjectSDK) -> Version | None:
-    parts = project_sdk.sdk_name.split(".", 2)
-    if len(parts) < 2:
-        return None
-
-    sdk_name = ".".join(parts[:2])
-
-    sdk_version_option = MINIMUM_SDK_VERSION_OPTIONS.get((project_sdk.event_type, sdk_name))
-    if sdk_version_option is None:
-        return None
-
-    sdk_version = options.get(sdk_version_option)
-    if sdk_version:
-        try:
-            return parse_version(sdk_version)
-        except InvalidVersion as e:
-            sentry_sdk.capture_exception(e)
-    return None
-
-
 def get_deprecation_status(project_sdk: ProjectSDK) -> SDKDeprecation | None:
     try:
         sdk_version = parse_version(project_sdk.sdk_version)
     except InvalidVersion as e:
         sentry_sdk.capture_exception(e)
         return None
 
-    minimum_sdk_version = get_minimum_sdk_version(project_sdk)
+    minimum_sdk_version = get_minimum_sdk_version(
+        project_sdk.event_type,
+        project_sdk.sdk_name,
+        hard_limit=False,
+    )
 
     # no minimum sdk version was specified
     if minimum_sdk_version is None:

src/sentry/features/temporary.py

@@ -273,6 +273,8 @@ def register_temporary_features(manager: FeatureManager):
     manager.add("organizations:profiling-beta", OrganizationFeature, FeatureHandlerStrategy.INTERNAL, api_expose=True)
     # Enables monitoring for latest profiling sdk used
     manager.add("organizations:profiling-sdks", OrganizationFeature, FeatureHandlerStrategy.FLAGPOLE, api_expose=True)
+    # Enables dropping of deprecated profiling sdks used
+    manager.add("organizations:profiling-deprecate-sdks", OrganizationFeature, FeatureHandlerStrategy.FLAGPOLE, api_expose=True)
     # Enables production profiling in sentry browser application
     manager.add("organizations:profiling-browser", OrganizationFeature, FeatureHandlerStrategy.INTERNAL, api_expose=False)
     # Enables separate differential flamegraph page

src/sentry/models/projectsdk.py

@@ -4,10 +4,12 @@
 from collections.abc import Sequence
 from enum import Enum
 
+import sentry_sdk
 from django.db import models
 from packaging.version import InvalidVersion, Version
 from packaging.version import parse as parse_version
 
+from sentry import options
 from sentry.backup.scopes import RelocationScope
 from sentry.db.models import BoundedIntegerField, FlexibleForeignKey, region_silo_model, sane_repr
 from sentry.db.models.base import DefaultFieldsModel
@@ -170,3 +172,33 @@ def normalize_sdk_name(sdk_name: str) -> str | None:
         return sdk_name
 
     return None
+
+
+MINIMUM_SDK_VERSION_OPTIONS: dict[tuple[int, str], str] = {
+    (EventType.PROFILE_CHUNK.value, "sentry.cocoa"): "sdk-deprecation.profile-chunk.cocoa",
+    (EventType.PROFILE_CHUNK.value, "sentry.python"): "sdk-deprecation.profile-chunk.python",
+}
+
+
+def get_minimum_sdk_version(event_type: int, sdk_name: str, hard_limit: bool) -> Version | None:
+    parts = sdk_name.split(".", 2)
+    if len(parts) < 2:
+        return None
+
+    normalized_sdk_name = ".".join(parts[:2])
+
+    sdk_version_option = MINIMUM_SDK_VERSION_OPTIONS.get((event_type, normalized_sdk_name))
+    if sdk_version_option is None:
+        return None
+
+    if hard_limit:
+        sdk_version = options.get(f"{sdk_version_option}.hard")
+    else:
+        sdk_version = options.get(sdk_version_option)
+
+    if sdk_version:
+        try:
+            return parse_version(sdk_version)
+        except InvalidVersion as e:
+            sentry_sdk.capture_exception(e)
+    return None

src/sentry/options/defaults.py

@@ -3194,8 +3194,18 @@
     default="2.24.1",
     flags=FLAG_AUTOMATOR_MODIFIABLE,
 )
+register(
+    "sdk-deprecation.profile-chunk.python.hard",
+    default="2.24.1",
+    flags=FLAG_AUTOMATOR_MODIFIABLE,
+)
 register(
     "sdk-deprecation.profile-chunk.cocoa",
+    default="8.49.2",
+    flags=FLAG_AUTOMATOR_MODIFIABLE,
+)
+register(
+    "sdk-deprecation.profile-chunk.cocoa.hard",
     default="8.49.0",
     flags=FLAG_AUTOMATOR_MODIFIABLE,
 )

src/sentry/profiles/task.py

@@ -15,6 +15,8 @@
 from arroyo import Topic as ArroyoTopic
 from arroyo.backends.kafka import KafkaPayload, KafkaProducer, build_kafka_configuration
 from django.conf import settings
+from packaging.version import InvalidVersion
+from packaging.version import parse as parse_version
 
 from sentry import features, options, quotas
 from sentry.conf.types.kafka_definition import Topic
@@ -27,7 +29,7 @@
 from sentry.models.files.utils import get_profiles_storage
 from sentry.models.organization import Organization
 from sentry.models.project import Project
-from sentry.models.projectsdk import EventType, ProjectSDK
+from sentry.models.projectsdk import EventType, ProjectSDK, get_minimum_sdk_version
 from sentry.profiles.java import (
     convert_android_methods_to_jvm_frames,
     deobfuscate_signature,
@@ -159,6 +161,45 @@ def process_profile_task(
     sentry_sdk.set_tag("project", project.id)
     sentry_sdk.set_tag("project.slug", project.slug)
 
+    if sampled:
+        if features.has("organizations:profiling-sdks", organization):
+            try:
+                event_type = determine_profile_type(profile)
+                sdk_name, sdk_version = determine_client_sdk(profile, event_type)
+
+                ProjectSDK.update_with_newest_version_or_create(
+                    project=project,
+                    event_type=event_type,
+                    sdk_name=sdk_name,
+                    sdk_version=sdk_version,
+                )
+
+                # Check to see if the data is coming from an deprecated SDK
+                # and drop it if needed
+                if is_sdk_deprecated(event_type, sdk_name, sdk_version):
+                    if features.has("organizations:profiling-deprecate-sdks", organization):
+                        category = (
+                            DataCategory.PROFILE_CHUNK
+                            if event_type == EventType.PROFILE_CHUNK
+                            else DataCategory.PROFILE
+                        )
+                        _track_outcome(
+                            profile=profile,
+                            project=project,
+                            outcome=Outcome.FILTERED,
+                            categories=[category],
+                            reason="deprecated sdk",
+                        )
+                        return
+            except UnableToAcquireLock:
+                # unable to acquire the lock means another event is trying to
+                # update the version so we can skip the update from this event
+                pass
+            except (UnknownClientSDKException, UnknownProfileTypeException):
+                pass
+            except Exception as e:
+                sentry_sdk.capture_exception(e)
+
     profile_context = {
         "organization_id": profile["organization_id"],
         "project_id": profile["project_id"],
@@ -236,14 +277,6 @@ def process_profile_task(
 
     if sampled:
         with metrics.timer("process_profile.track_outcome.accepted"):
-            if features.has("organizations:profiling-sdks", organization):
-                try:
-                    track_latest_sdk(project, profile)
-                except (UnknownClientSDKException, UnknownProfileTypeException):
-                    pass
-                except Exception as e:
-                    sentry_sdk.capture_exception(e)
-
             if not project.flags.has_profiles:
                 first_profile_received.send_robust(project=project, sender=Project)
             try:
@@ -1212,21 +1245,32 @@ def determine_client_sdk(profile: Profile, event_type: EventType) -> tuple[str,
     raise UnknownClientSDKException
 
 
-def track_latest_sdk(project: Project, profile: Profile) -> None:
-    event_type = determine_profile_type(profile)
-    sdk_name, sdk_version = determine_client_sdk(profile, event_type)
+def is_sdk_deprecated(event_type: EventType, sdk_name: str, sdk_version: str) -> bool:
+    minimum_version = get_minimum_sdk_version(event_type.value, sdk_name, hard_limit=True)
+
+    # no minimum sdk version was specified
+    if minimum_version is None:
+        return False
 
     try:
-        ProjectSDK.update_with_newest_version_or_create(
-            project=project,
-            event_type=event_type,
-            sdk_name=sdk_name,
-            sdk_version=sdk_version,
+        version = parse_version(sdk_version)
+    except InvalidVersion:
+        return False
+
+    # satisfies the minimum sdk version
+    if version >= minimum_version:
+        return False
+
+    parts = sdk_name.split(".", 2)
+    if len(parts) >= 2:
+        normalized_sdk_name = ".".join(parts[:2])
+        metrics.incr(
+            "process_profile.sdk.deprecated",
+            tags={"sdk_name": normalized_sdk_name},
+            sample_rate=1.0,
         )
-    except UnableToAcquireLock:
-        # unable to acquire the lock means another event is trying to update the version
-        # so we can skip the update from this event
-        pass
+
+    return True
 
 
 @metrics.wraps("process_profile.process_vroomrs_profile")

tests/sentry/api/endpoints/test_organization_sdk_deprecations.py

@@ -139,7 +139,7 @@ def test_mixed_sdks(self):
             project=self.project,
             event_type=EventType.PROFILE_CHUNK.value,
             sdk_name="sentry.cocoa",
-            sdk_version="8.49.0",
+            sdk_version="8.49.2",
         )
         response = self.client.get(
             self.url,

tests/sentry/profiles/test_task.py

@@ -33,7 +33,7 @@
 from sentry.profiles.utils import Profile
 from sentry.testutils.cases import TransactionTestCase
 from sentry.testutils.factories import Factories, get_fixture_path
-from sentry.testutils.helpers import Feature
+from sentry.testutils.helpers import Feature, override_options
 from sentry.testutils.pytest.fixtures import django_db_all
 from sentry.testutils.skips import requires_symbolicator
 from sentry.utils import json
@@ -206,7 +206,7 @@ def sample_v1_profile():
   },
   "client_sdk": {
     "name": "sentry.python",
-    "version": "2.23.0"
+    "version": "2.24.1"
   }
 }"""
     )
@@ -273,7 +273,7 @@ def generate_sample_v2_profile():
   },
   "client_sdk": {
     "name": "sentry.python",
-    "version": "2.23.0"
+    "version": "2.24.1"
   }
 }"""
     )
@@ -976,7 +976,7 @@ def test_track_latest_sdk(
             project=project,
             event_type=event_type.value,
             sdk_name="sentry.python",
-            sdk_version="2.23.0",
+            sdk_version="2.24.1",
         )
         is not None
     )
@@ -1079,7 +1079,65 @@ def test_track_latest_sdk_with_payload(
             project=project,
             event_type=EventType.PROFILE.value,
             sdk_name="sentry.python",
-            sdk_version="2.23.0",
+            sdk_version="2.24.1",
         )
         is not None
     )
+
+
+@patch("sentry.profiles.task._track_outcome")
+@patch("sentry.profiles.task._push_profile_to_vroom")
+@django_db_all
+@pytest.mark.parametrize(
+    ["profile", "category", "sdk_version", "dropped"],
+    [
+        pytest.param("sample_v1_profile", DataCategory.PROFILE, "2.23.0", False),
+        pytest.param("sample_v2_profile", DataCategory.PROFILE_CHUNK, "2.23.0", True),
+        pytest.param("sample_v2_profile", DataCategory.PROFILE_CHUNK, "2.24.0", False),
+        pytest.param("sample_v2_profile", DataCategory.PROFILE_CHUNK, "2.24.1", False),
+    ],
+)
+def test_deprecated_sdks(
+    _push_profile_to_vroom,
+    _track_outcome,
+    profile,
+    category,
+    sdk_version,
+    dropped,
+    organization,
+    project,
+    request,
+):
+    profile = request.getfixturevalue(profile)
+    profile["organization_id"] = organization.id
+    profile["project_id"] = project.id
+    profile["client_sdk"] = {
+        "name": "sentry.python",
+        "version": sdk_version,
+    }
+
+    with Feature(
+        [
+            "organizations:profiling-sdks",
+            "organizations:profiling-deprecate-sdks",
+        ]
+    ):
+        with override_options(
+            {
+                "sdk-deprecation.profile-chunk.python": "2.24.1",
+                "sdk-deprecation.profile-chunk.python.hard": "2.24.0",
+            }
+        ):
+            process_profile_task(profile=profile)
+
+    if dropped:
+        _push_profile_to_vroom.assert_not_called()
+        _track_outcome.assert_called_with(
+            profile=profile,
+            project=project,
+            outcome=Outcome.FILTERED,
+            categories=[category],
+            reason="deprecated sdk",
+        )
+    else:
+        _push_profile_to_vroom.assert_called()