Bug 1700606 - Expire garbage backfill reports

* also remove alerts older than a year
* slightly refactor main garbage collection method
* fix ORM filter
* provide test coverage
* reduce ambiguity around signature garbage collection

tests/model/cycle_data/test_perfherder_cycling.py

@@ -24,9 +24,15 @@ from treeherder.perf.models import (
     PerformanceAlertSummary,
     PerformanceAlert,
     MultiCommitDatum,
+    BackfillReport,
 )
 
 
+@pytest.fixture
+def empty_backfill_report(test_perf_alert_summary) -> BackfillReport:
+    return BackfillReport.objects.create(summary=test_perf_alert_summary)
+
+
 @pytest.mark.parametrize(
     'repository_name',
     [
@@ -840,3 +846,34 @@ def test_deleting_performance_data_cascades_to_perf_multicomit_data(test_perf_da
         cursor.close()
 
     assert MultiCommitDatum.objects.count() == 0
+
+
+def test_alerts_older_than_a_year_are_removed_even_if_signature_is_active(test_perf_alert):
+    # alert (fixture) comes pre linked to an active signature
+    test_perf_alert.created = datetime.now() - timedelta(days=365)
+    test_perf_alert.save()
+
+    PerfherderCycler(10_000, 0).cycle()
+
+    assert not PerformanceAlert.objects.filter(id=test_perf_alert.id).exists()
+
+
+def test_empty_backfill_reports_get_removed(empty_backfill_report):
+    empty_backfill_report.created = datetime.now() - timedelta(days=120)
+    empty_backfill_report.save()
+
+    PerfherderCycler(10_000, 0).cycle()
+
+    assert BackfillReport.objects.count() == 0
+
+
+@pytest.mark.parametrize('days_since_created', [0, 30, 100])
+def test_empty_backfill_reports_arent_removed_if_not_enough_time_passed(
+    empty_backfill_report, days_since_created
+):
+    empty_backfill_report.created = datetime.now() - timedelta(days=days_since_created)
+    empty_backfill_report.save()
+
+    PerfherderCycler(10_000, 0).cycle()
+
+    assert BackfillReport.objects.filter(summary_id=empty_backfill_report.summary_id).exists()

treeherder/model/data_cycling/cyclers.py

@@ -10,7 +10,12 @@ from django.db.models import Count
 from treeherder.model.data_cycling.removal_strategies import RemovalStrategy
 from treeherder.model.models import Job, JobType, JobGroup, Machine
 from treeherder.perf.exceptions import NoDataCyclingAtAll, MaxRuntimeExceeded
-from treeherder.perf.models import PerformanceSignature, PerformanceAlertSummary
+from treeherder.perf.models import (
+    PerformanceSignature,
+    PerformanceAlertSummary,
+    PerformanceAlert,
+    BackfillReport,
+)
 from treeherder.services import taskcluster
 from .max_runtime import MaxRuntime
 from .signature_remover import PublicSignatureRemover
@@ -130,15 +135,35 @@ class PerfherderCycler(DataCycler):
             logger.warning(ex)
 
     def _remove_leftovers(self):
-        # remove any signatures which are
-        # no longer associated with a job
-        signatures = PerformanceSignature.objects.filter(last_updated__lte=self.max_timestamp)
-        notify_client = taskcluster.notify_client_factory()
-        signatures_remover = PublicSignatureRemover(timer=self.timer, notify_client=notify_client)
-        signatures_remover.remove_in_chunks(signatures)
+        self.__remove_empty_signatures()
 
-        # remove empty alert summaries
-        logger.warning('Removing alert summaries which no longer have any alerts...')
+        self.__remove_too_old_alerts()
+        self.__remove_empty_alert_summaries()
+
+        self.__remove_empty_backfill_reports()
+
+    def __remove_empty_signatures(self):
+        logger.warning("Removing performance signatures which don't have any data points...")
+        potentially_empty_signatures = PerformanceSignature.objects.filter(
+            last_updated__lte=self.max_timestamp
+        )
+        notify_client = taskcluster.notify_client_factory()
+
+        signatures_remover = PublicSignatureRemover(timer=self.timer, notify_client=notify_client)
+        signatures_remover.remove_in_chunks(potentially_empty_signatures)
+
+    def __remove_too_old_alerts(self):
+        logger.warning("Removing alerts older than a year...")
+        PerformanceAlert.objects.filter(
+            # WARNING! Don't change this without proper approval!           #
+            # Otherwise we risk deleting data that's actively investigated  #
+            # and cripple the perf sheriffing process!                      #
+            created__lt=(datetime.now() - timedelta(days=365))
+            #################################################################
+        ).delete()
+
+    def __remove_empty_alert_summaries(self):
+        logger.warning("Removing alert summaries which no longer have any alerts...")
         (
             PerformanceAlertSummary.objects.prefetch_related('alerts', 'related_alerts')
             .annotate(
@@ -157,6 +182,14 @@ class PerfherderCycler(DataCycler):
             .delete()
         )
 
+    def __remove_empty_backfill_reports(self):
+        logger.warning("Removing backfill reports which no longer have any records...")
+        four_months_ago = datetime.now() - timedelta(days=120)
+
+        BackfillReport.objects.annotate(total_records=Count('records')).filter(
+            created__lt=four_months_ago, total_records=0
+        ).delete()
+
     def _delete_in_chunks(self, strategy: RemovalStrategy):
         any_successful_attempt = False
 

treeherder/model/data_cycling/signature_remover.py

@@ -4,6 +4,7 @@ from typing import List
 import taskcluster
 from django.conf import settings
 from django.db import transaction
+from django.db.models import QuerySet
 from taskcluster.exceptions import TaskclusterRestFailure
 
 from treeherder.perf.models import PerformanceSignature
@@ -34,14 +35,13 @@ class PublicSignatureRemover:
         self._email_writer = email_writer or DeletionNotificationWriter()
         self.timer = timer
 
-    def remove_in_chunks(self, signatures):
+    def remove_in_chunks(self, potentially_empty_signatures: QuerySet):
         emails_sent = 0
         rows_left = self._max_rows_allowed
         chunk_of_signatures = []
 
-        logger.warning("Removing performance signatures which don't have any data points...")
-        self._remove_empty_try_signatures(signatures)
-        for perf_signature in signatures:
+        self._remove_empty_try_signatures(potentially_empty_signatures)
+        for perf_signature in potentially_empty_signatures:
             self.timer.quit_on_timeout()
 
             if emails_sent < self._max_emails_allowed and (
@@ -63,7 +63,7 @@ class PublicSignatureRemover:
             self.__delete_and_notify(chunk_of_signatures)
 
     @staticmethod
-    def _remove_empty_try_signatures(signatures):
+    def _remove_empty_try_signatures(signatures: QuerySet):
         try_signatures = signatures.filter(repository__name='try')
         for perf_signature in try_signatures:
             if not perf_signature.has_performance_data():