Merge pull request #35025 from dimagi/mjr/disable-sentry-dedupe-error

Ignore failing dedupe records with long stale ES data

corehq/apps/data_interfaces/models.py

@@ -1158,7 +1158,13 @@ def _handle_case_duplicate(self, case, rule):
                 # We've decided skipping this record and recording an error is likely the safest way to handle this
                 # Hopefully, these errors allow us to track down the underlying bug or infrastructure issue
                 # and fix the issue at the source
-                raise ValueError(f'Unable to find current ElasticSearch data for: {case.case_id}')
+
+                # TODO: Commenting this line out, as it has caused us to surpass our sentry error quota
+                # we should figure out whether this can be safely handled, or what truly constitutes an error,
+                # but disabling this to avoid further quota issues.
+                # raise ValueError(f'Unable to find current ElasticSearch data for: {case.case_id}')
+                # Ignore this result for now
+                return CaseRuleActionResult(num_errors=1)
             else:
                 # Normal processing can involve latency between when a case is written to the database and when
                 # it arrives in ElasticSearch. If this case was modified within the acceptable latency window,

corehq/apps/data_interfaces/tests/test_case_deduplication.py

@@ -642,16 +642,15 @@ def test_ignores_recent_submissions_not_in_elasticsearch(self, mock_case_exists,
         self.assertEqual(created_duplicates.count(), 0)
 
     @patch("corehq.apps.data_interfaces.models.case_matching_rule_criteria_exists_in_es")
-    def test_raises_error_when_case_not_in_elasticsearch(self, mock_case_exists):
+    def test_returns_error_result_when_case_not_in_elasticsearch(self, mock_case_exists):
         current_time = datetime(year=2024, month=2, day=5, hour=10)
         case = self._create_case()
         case.server_modified_on = current_time - timedelta(hours=1, seconds=1)
         mock_case_exists.return_value = False
 
-        with self.assertRaisesRegex(
-                ValueError, f'Unable to find current ElasticSearch data for: {case.case_id}'):
-            with freeze_time(current_time):
-                self.rule.run_rule(case, datetime.now())
+        result = self.rule.run_rule(case, datetime.now())
+        self.assertEqual(result.num_updates, 0)
+        self.assertEqual(result.num_errors, 1)
 
     @patch("corehq.apps.data_interfaces.models._find_duplicate_case_ids")
     def test_case_no_longer_duplicate(self, find_duplicates_mock):