Merge pull request #5108 from mozilla/metrics_event_no_ga_mpp_3831

MPP-3831: Send metrics_event data to influxdb, logs instead of UA

METRICS.md

@@ -1492,19 +1492,21 @@ The Google Analytics events:
 The Relay Extension makes it easier to use Relay email masks on websites. It is
 available as a [Firefox Extension][] and a [Chrome Extension][].
 
+The Relay Extension uses a background listener to send interaction events to
+the Relay API server. These events are then recorded as statsd-style statistics
+and in server logs.
+
+The Relay Extension generates a [random UUID][] for the extension identifier
+that is [stored locally in the browser][]. A different ID will be generated for
+each browser and machine. A truncated hash of this identifier is included in the
+system logs to estimate the count of unique extension installations.
+
 <!-- References in this paragraph are defined in section "Google Analytics" -->
 
-The Relay Extension uses a background listener to send interaction events to
-the Relay API server. The API server forwards the events to Google Analytics.
-Events are reported using the [Universal Measurement Protocol][].
+Before October 2024, the API server forwarded the events
+to Google Analytics, using the [Universal Measurement Protocol][].
 Google [replaced Universal Analytics with Google Analytics 4][] (GA4) on July 1, 2024,
-and these events are no longer recorded. Relay is in the process of switching
-to [GA4][].
-
-The Relay Extension generates a [random UUID][] for the Google Analytics
-identifier that is [stored locally in the browser][]. This ID is different from
-the GA identifier on the Relay webpage. A different ID will be generated for
-each browser and machine.
+and these events stopped being recorded.
 
 [Chrome Extension]: https://chromewebstore.google.com/detail/firefox-relay/lknpoadjjkjcmjhbjpcljdednccbldeb "The Firefox Relay extension on the Chrome Web Store"
 [Firefox Extension]: https://addons.mozilla.org/en-US/firefox/addon/private-relay/ "The Firefox Relay extension on Firefox Browser Add-Ons"

privaterelay/tests/views_tests.py

@@ -1,16 +1,16 @@
 import json
 import logging
-from collections.abc import Callable, Iterator
+from collections.abc import Iterator
 from copy import deepcopy
 from dataclasses import dataclass
 from datetime import UTC, datetime, timedelta
 from pathlib import Path
 from typing import Any, Literal
-from unittest.mock import ANY, Mock, patch
+from unittest.mock import Mock, patch
 from uuid import uuid4
 
 from django.contrib.auth.models import User
-from django.test import Client, TestCase
+from django.test import Client, TestCase, override_settings
 from django.utils import timezone
 
 import jwt
@@ -38,7 +38,7 @@
 from ..apps import PrivateRelayConfig
 from ..fxa_utils import NoSocialToken
 from ..models import Profile
-from ..views import _update_all_data, fxa_verifying_keys, send_ga_ping
+from ..views import _update_all_data, fxa_verifying_keys
 
 
 def test_no_social_token():
@@ -598,82 +598,59 @@ def test_lbheartbeat_view(client: Client) -> None:
     assert response.content == b""
 
 
-_ThreadAndReportMocks = dict[Literal["thread", "report"], Mock]
-
-
-@pytest.fixture
-def mock_metrics_thread_and_report() -> Iterator[_ThreadAndReportMocks]:
-    """
-    Setup mocks for metrics event.
-
-    Replace google_measurement_protocol.report with a Mock
-    Replace Thread with a mock version that calls immediately.
-    """
-
-    with (
-        patch("privaterelay.views.threading.Thread", spec=True) as mock_thread_cls,
-        patch("privaterelay.views.report") as mock_report,
-    ):
-
-        mock_thread = Mock(spec_set=["start"])
-
-        def create_thread(
-            target: Callable[[str, str, Any], None],
-            args: tuple[str, str, Any],
-            daemon: bool,
-        ) -> Mock:
-            assert target == send_ga_ping
-            assert daemon
-
-            def call_send_ga_ping() -> None:
-                target(*args)
-
-            mock_thread.start.side_effect = call_send_ga_ping
-            return mock_thread
-
-        mock_thread_cls.side_effect = create_thread
-        yield {"thread": mock_thread, "report": mock_report}
-
-
-def test_metrics_event_GET(
-    client: Client, mock_metrics_thread_and_report: _ThreadAndReportMocks
-) -> None:
-    response = client.get("/metrics-event")
+@override_settings(STATSD_ENABLED=True)
+def test_metrics_event_GET(client: Client, caplog: pytest.LogCaptureFixture) -> None:
+    with MetricsMock() as mm:
+        response = client.get("/metrics-event")
     assert response.status_code == 405
-    mock_metrics_thread_and_report["thread"].start.assert_not_called()
-    mock_metrics_thread_and_report["report"].assert_not_called()
+    assert caplog.record_tuples == [("request.summary", logging.INFO, "")]
+    mm.assert_not_incr("fx.private.relay.metrics_event")
 
 
+@override_settings(STATSD_ENABLED=True)
 def test_metrics_event_POST_non_json(
-    client: Client, mock_metrics_thread_and_report: _ThreadAndReportMocks
+    client: Client, caplog: pytest.LogCaptureFixture
 ) -> None:
-    response = client.post("/metrics-event")
+    with MetricsMock() as mm:
+        response = client.post("/metrics-event")
     assert response.status_code == 415
-    mock_metrics_thread_and_report["thread"].start.assert_not_called()
-    mock_metrics_thread_and_report["report"].assert_not_called()
+    assert caplog.record_tuples == [("request.summary", logging.INFO, "")]
+    mm.assert_not_incr("fx.private.relay.metrics_event")
 
 
+@override_settings(STATSD_ENABLED=True)
 def test_metrics_event_POST_json_no_ga_uuid(
-    client: Client, mock_metrics_thread_and_report: _ThreadAndReportMocks
+    client: Client, caplog: pytest.LogCaptureFixture
 ) -> None:
-    response = client.post(
-        "/metrics-event", {"category": "addon"}, content_type="application/json"
-    )
+    with MetricsMock() as mm:
+        response = client.post(
+            "/metrics-event", {"category": "addon"}, content_type="application/json"
+        )
     assert response.status_code == 404
-    mock_metrics_thread_and_report["thread"].start.assert_not_called()
-    mock_metrics_thread_and_report["report"].assert_not_called()
+    assert caplog.record_tuples == [("request.summary", logging.INFO, "")]
+    mm.assert_not_incr("fx.private.relay.metrics_event")
 
 
+@override_settings(STATSD_ENABLED=True)
 def test_metrics_event_POST_json_ga_uuid_ok(
     client: Client,
-    mock_metrics_thread_and_report: _ThreadAndReportMocks,
+    caplog: pytest.LogCaptureFixture,
     settings: SettingsWrapper,
 ) -> None:
-    response = client.post(
-        "/metrics-event", {"ga_uuid": "anything-is-ok"}, content_type="application/json"
-    )
+    with MetricsMock() as mm:
+        response = client.post(
+            "/metrics-event",
+            {"ga_uuid": "anything-is-ok"},
+            content_type="application/json",
+        )
     assert response.status_code == 200
-    mock_metrics_thread_and_report["thread"].start.assert_called_once_with()
-    mock_metrics_thread_and_report["report"].assert_called_once_with(
-        settings.GOOGLE_ANALYTICS_ID, "anything-is-ok", ANY
-    )
+
+    assert caplog.record_tuples == [
+        ("eventsinfo", logging.INFO, "metrics_event"),
+        ("request.summary", logging.INFO, ""),
+    ]
+    record = caplog.records[0]
+    assert getattr(record, "ga_uuid_hash") == "1aa8606ede8415d8"
+    assert getattr(record, "source") == "website"
+
+    mm.assert_incr_once("fx.private.relay.metrics_event", 1, tags=["source:website"])

privaterelay/views.py

@@ -1,6 +1,5 @@
 import json
 import logging
-import threading
 from collections.abc import Iterable
 from datetime import UTC, datetime
 from functools import cache
@@ -21,7 +20,7 @@
 from allauth.socialaccount.models import SocialAccount, SocialApp
 from allauth.socialaccount.providers.fxa.views import FirefoxAccountsOAuth2Adapter
 from cryptography.hazmat.primitives.asymmetric.rsa import RSAPublicKey
-from google_measurement_protocol import event, report
+from markus.utils import generate_tag
 from oauthlib.oauth2.rfc6749.errors import CustomOAuth2Error
 from rest_framework.decorators import api_view, schema
 
@@ -91,40 +90,38 @@ def profile_subdomain(request):
         return JsonResponse({"message": e.message, "subdomain": subdomain}, status=400)
 
 
-def send_ga_ping(ga_id: str, ga_uuid: str, data: Any) -> None:
-    try:
-        report(ga_id, ga_uuid, data)
-    except Exception as e:
-        logger.error("metrics_event", extra={"error": e})
-
-
 @csrf_exempt
 @require_http_methods(["POST"])
 def metrics_event(request: HttpRequest) -> JsonResponse:
+    """
+    Handle metrics events from the Relay extension.
+
+    This used to forward data to Google Analytics, but was not updated for GA4.
+
+    Now it logs the information and updates statsd counters.
+    """
     try:
         request_data = json.loads(request.body)
     except json.JSONDecodeError:
         return JsonResponse({"msg": "Could not decode JSON"}, status=415)
     if "ga_uuid" not in request_data:
         return JsonResponse({"msg": "No GA uuid found"}, status=404)
-    # "dimension5" is a Google Analytics-specific variable to track a custom dimension,
-    # used to determine which browser vendor the add-on is using: Firefox or Chrome
-    # "dimension7" is a Google Analytics-specific variable to track a custom dimension,
-    # used to determine where the ping is coming from: website (default), add-on or app
-    event_data = event(
-        request_data.get("category", None),
-        request_data.get("action", None),
-        request_data.get("label", None),
-        request_data.get("value", None),
-        dimension5=request_data.get("dimension5", None),
-        dimension7=request_data.get("dimension7", "website"),
-    )
-    t = threading.Thread(
-        target=send_ga_ping,
-        args=[settings.GOOGLE_ANALYTICS_ID, request_data.get("ga_uuid"), event_data],
-        daemon=True,
-    )
-    t.start()
+    event_data = {
+        "ga_uuid_hash": sha256(request_data["ga_uuid"].encode()).hexdigest()[:16],
+        "category": request_data.get("category", None),
+        "action": request_data.get("action", None),
+        "label": request_data.get("label", None),
+        "value": request_data.get("value", None),
+        "browser": request_data.get("browser", None),  # dimension5 in GA
+        "source": request_data.get("dimension7", "website"),
+    }
+    info_logger.info("metrics_event", extra=event_data)
+    tags = [
+        generate_tag(key, val)
+        for key, val in event_data.items()
+        if val is not None and key != "ga_uuid_hash"
+    ]
+    incr_if_enabled("metrics_event", tags=tags)
     return JsonResponse({"msg": "OK"}, status=200)
 
 

requirements.txt

@@ -19,7 +19,6 @@ dockerflow==2024.4.2
 drf-spectacular==0.27.2
 drf-spectacular-sidecar==2024.7.1
 glean_parser==15.0.1
-google-measurement-protocol==1.1.0
 google-cloud-profiler==4.1.0
 google-cloud-sqlcommenter==2.0.0; python_version < '3.12'
 gunicorn==23.0.0