fix(api): resolve check_title filter to check_id for consistent finding-group counts (#10486)

Author: AdriiiPRodri

api/CHANGELOG.md

@@ -16,6 +16,7 @@ All notable changes to the **Prowler API** are documented in this file.
 - Filter transient Neo4j defunct connection logs in Sentry `before_send` to suppress false-positive alerts handled by `RetryableSession` retries [(#10452)](https://github.com/prowler-cloud/prowler/pull/10452)
 - `MANAGE_ACCOUNT` permission no longer required for listing and creating tenants [(#10468)](https://github.com/prowler-cloud/prowler/pull/10468)
 - Finding groups muted filter, counters, metadata extraction and mute reaggregation [(#10477)](https://github.com/prowler-cloud/prowler/pull/10477)
+- Finding groups `check_title__icontains` resolution, `name__icontains` resource filter and `resource_group` field in `/resources` response [(#10486)](https://github.com/prowler-cloud/prowler/pull/10486)
 
 ## [1.23.0] (Prowler v5.22.0)
 

api/src/backend/api/filters.py

@@ -44,6 +44,7 @@
     ProviderGroup,
     ProviderSecret,
     Resource,
+    ResourceFindingMapping,
     ResourceTag,
     Role,
     Scan,
@@ -197,17 +198,13 @@ class CommonFindingFilters(FilterSet):
         field_name="resource_services", lookup_expr="icontains"
     )
 
-    resource_uid = CharFilter(field_name="resources__uid")
-    resource_uid__in = CharInFilter(field_name="resources__uid", lookup_expr="in")
-    resource_uid__icontains = CharFilter(
-        field_name="resources__uid", lookup_expr="icontains"
-    )
+    resource_uid = CharFilter(method="filter_resource_uid")
+    resource_uid__in = CharInFilter(method="filter_resource_uid_in")
+    resource_uid__icontains = CharFilter(method="filter_resource_uid_icontains")
 
-    resource_name = CharFilter(field_name="resources__name")
-    resource_name__in = CharInFilter(field_name="resources__name", lookup_expr="in")
-    resource_name__icontains = CharFilter(
-        field_name="resources__name", lookup_expr="icontains"
-    )
+    resource_name = CharFilter(method="filter_resource_name")
+    resource_name__in = CharInFilter(method="filter_resource_name_in")
+    resource_name__icontains = CharFilter(method="filter_resource_name_icontains")
 
     resource_type = CharFilter(method="filter_resource_type")
     resource_type__in = CharInFilter(field_name="resource_types", lookup_expr="overlap")
@@ -266,10 +263,49 @@ def filter_resource_tag(self, queryset, name, value):
         return queryset.filter(overall_query).distinct()
 
     def filter_check_title_icontains(self, queryset, name, value):
+        # Resolve from the summary table (has check_title column + trigram
+        # GIN index) instead of scanning JSON in the findings table.
+        matching_check_ids = (
+            FindingGroupDailySummary.objects.filter(
+                check_title__icontains=value,
+            )
+            .values_list("check_id", flat=True)
+            .distinct()
+        )
+        return queryset.filter(check_id__in=matching_check_ids)
+
+    # --- Resource subquery filters ---
+    # Resolve resource → RFM → finding_ids first, then filter findings
+    # by id__in.  This avoids a 3-way JOIN driven from the (huge)
+    # findings side and lets PostgreSQL start from the resources
+    # unique-constraint index instead.
+
+    @staticmethod
+    def _finding_ids_for_resources(**lookup):
+        return ResourceFindingMapping.objects.filter(
+            resource__in=Resource.objects.filter(**lookup).values("id")
+        ).values("finding_id")
+
+    def filter_resource_uid(self, queryset, name, value):
+        return queryset.filter(id__in=self._finding_ids_for_resources(uid=value))
+
+    def filter_resource_uid_in(self, queryset, name, value):
+        return queryset.filter(id__in=self._finding_ids_for_resources(uid__in=value))
+
+    def filter_resource_uid_icontains(self, queryset, name, value):
         return queryset.filter(
-            Q(check_metadata__CheckTitle__icontains=value)
-            | Q(check_metadata__checktitle__icontains=value)
-            | Q(check_metadata__Checktitle__icontains=value)
+            id__in=self._finding_ids_for_resources(uid__icontains=value)
+        )
+
+    def filter_resource_name(self, queryset, name, value):
+        return queryset.filter(id__in=self._finding_ids_for_resources(name=value))
+
+    def filter_resource_name_in(self, queryset, name, value):
+        return queryset.filter(id__in=self._finding_ids_for_resources(name__in=value))
+
+    def filter_resource_name_icontains(self, queryset, name, value):
+        return queryset.filter(
+            id__in=self._finding_ids_for_resources(name__icontains=value)
         )
 
 
@@ -919,7 +955,19 @@ class Meta:
         }
 
 
-class FindingGroupSummaryFilter(FilterSet):
+class _CheckTitleToCheckIdMixin:
+    """Resolve check_title search to check_ids so all provider rows are kept."""
+
+    def filter_check_title_to_check_ids(self, queryset, name, value):
+        matching_check_ids = (
+            queryset.filter(check_title__icontains=value)
+            .values_list("check_id", flat=True)
+            .distinct()
+        )
+        return queryset.filter(check_id__in=matching_check_ids)
+
+
+class FindingGroupSummaryFilter(_CheckTitleToCheckIdMixin, FilterSet):
     """
     Filter for FindingGroupDailySummary queries.
 
@@ -942,9 +990,7 @@ class FindingGroupSummaryFilter(FilterSet):
     check_id = CharFilter(field_name="check_id", lookup_expr="exact")
     check_id__in = CharInFilter(field_name="check_id", lookup_expr="in")
     check_id__icontains = CharFilter(field_name="check_id", lookup_expr="icontains")
-    check_title__icontains = CharFilter(
-        field_name="check_title", lookup_expr="icontains"
-    )
+    check_title__icontains = CharFilter(method="filter_check_title_to_check_ids")
 
     # Provider filters
     provider_id = UUIDFilter(field_name="provider_id", lookup_expr="exact")
@@ -1032,7 +1078,7 @@ def _maybe_date_to_datetime(value):
         return dt
 
 
-class LatestFindingGroupSummaryFilter(FilterSet):
+class LatestFindingGroupSummaryFilter(_CheckTitleToCheckIdMixin, FilterSet):
     """
     Filter for FindingGroupDailySummary /latest endpoint.
 
@@ -1044,9 +1090,7 @@ class LatestFindingGroupSummaryFilter(FilterSet):
     check_id = CharFilter(field_name="check_id", lookup_expr="exact")
     check_id__in = CharInFilter(field_name="check_id", lookup_expr="in")
     check_id__icontains = CharFilter(field_name="check_id", lookup_expr="icontains")
-    check_title__icontains = CharFilter(
-        field_name="check_title", lookup_expr="icontains"
-    )
+    check_title__icontains = CharFilter(method="filter_check_title_to_check_ids")
 
     # Provider filters
     provider_id = UUIDFilter(field_name="provider_id", lookup_expr="exact")

api/src/backend/api/tests/test_views.py

@@ -15810,6 +15810,50 @@ def test_finding_groups_check_title_icontains(
         assert len(data) == 1
         assert data[0]["id"] == "s3_bucket_public_access"
 
+    @pytest.mark.parametrize(
+        "extra_filters",
+        [
+            {},
+            {"filter[muted]": "include"},
+        ],
+        ids=["summary_path", "finding_level_path"],
+    )
+    def test_check_title_icontains_includes_all_title_variants(
+        self,
+        authenticated_client,
+        finding_groups_title_variants_fixture,
+        extra_filters,
+    ):
+        """
+        Regression: two providers report the same check_id with different
+        checktitle values (e.g. after a Prowler version upgrade).  Filtering
+        by check_title__icontains with a term that matches only ONE variant
+        must still return the finding group with counts from BOTH providers.
+
+        Parametrized to cover both aggregation paths:
+        - summary_path: default, uses _CheckTitleToCheckIdMixin on summaries
+        - finding_level_path: filter[muted]=include forces CommonFindingFilters
+        """
+        params = {
+            "filter[inserted_at]": TODAY,
+            "filter[check_title.icontains]": "Ensure repository",
+            **extra_filters,
+        }
+        response = authenticated_client.get(
+            reverse("finding-group-list"),
+            params,
+        )
+        assert response.status_code == status.HTTP_200_OK
+        data = response.json()["data"]
+        assert len(data) == 1
+        assert data[0]["id"] == "github_secret_scanning_enabled"
+        attrs = data[0]["attributes"]
+        # Both providers' findings must be counted
+        assert attrs["fail_count"] == 2, (
+            "fail_count must include findings from both providers, "
+            "regardless of which title variant matches the search"
+        )
+
     def test_resources_not_found(self, authenticated_client):
         """Test 404 returned for nonexistent check_id."""
         response = authenticated_client.get(
@@ -15851,6 +15895,44 @@ def test_resources_fields(self, authenticated_client, finding_groups_fixture):
             assert resource.get("region"), "resource.region must not be empty"
             assert resource.get("type"), "resource.type must not be empty"
 
+    def test_resources_resource_group(
+        self, authenticated_client, finding_groups_fixture
+    ):
+        """Test resource_group is extracted from check_metadata.resourcegroup."""
+        response = authenticated_client.get(
+            reverse(
+                "finding-group-resources", kwargs={"pk": "s3_bucket_public_access"}
+            ),
+            {"filter[inserted_at]": TODAY},
+        )
+        assert response.status_code == status.HTTP_200_OK
+        data = response.json()["data"]
+        assert len(data) == 2
+        for item in data:
+            resource = item["attributes"]["resource"]
+            assert (
+                resource["resource_group"] == "storage"
+            ), "resource_group must be 'storage'"
+
+    def test_resources_name_icontains(
+        self, authenticated_client, finding_groups_fixture
+    ):
+        """Test resource_name__icontains filters resources by name substring."""
+        # s3_bucket_public_access has "My Instance 1" and "My Instance 2"
+        response = authenticated_client.get(
+            reverse(
+                "finding-group-resources", kwargs={"pk": "s3_bucket_public_access"}
+            ),
+            {
+                "filter[inserted_at]": TODAY,
+                "filter[resource_name.icontains]": "Instance 1",
+            },
+        )
+        assert response.status_code == status.HTTP_200_OK
+        data = response.json()["data"]
+        assert len(data) == 1
+        assert "Instance 1" in data[0]["attributes"]["resource"]["name"]
+
     def test_resources_provider_info(
         self, authenticated_client, finding_groups_fixture
     ):

api/src/backend/api/v1/serializers.py

@@ -4194,6 +4194,7 @@ class JSONAPIMeta:
                 "service": {"type": "string"},
                 "region": {"type": "string"},
                 "type": {"type": "string"},
+                "resource_group": {"type": "string"},
             },
         }
     )
@@ -4205,6 +4206,7 @@ def get_resource(self, obj):
             "service": obj.get("resource_service", ""),
             "region": obj.get("resource_region", ""),
             "type": obj.get("resource_type", ""),
+            "resource_group": obj.get("resource_group", ""),
         }
 
     @extend_schema_field(

api/src/backend/api/v1/views.py

@@ -3483,7 +3483,7 @@ def list(self, request, *args, **kwargs):
             request,
             filtered_queryset,
             manager=Finding.all_objects,
-            select_related=["scan"],
+            select_related=["scan__provider"],
             prefetch_related=["resources"],
         )
 
@@ -3653,7 +3653,7 @@ def latest(self, request):
         tenant_id = request.tenant_id
         filtered_queryset = self.filter_queryset(self.get_queryset())
 
-        latest_scan_ids = (
+        latest_scan_ids = list(
             Scan.all_objects.filter(tenant_id=tenant_id, state=StateChoices.COMPLETED)
             .order_by("provider_id", "-inserted_at")
             .distinct("provider_id")
@@ -3667,7 +3667,7 @@ def latest(self, request):
             request,
             filtered_queryset,
             manager=Finding.all_objects,
-            select_related=["scan"],
+            select_related=["scan__provider"],
             prefetch_related=["resources"],
         )
 
@@ -6878,8 +6878,15 @@ def retrieve(self, request, *args, **kwargs):
         "resource_type__icontains": "type__icontains",
     }
 
+    # Fields accepted directly by LatestResourceFilter (no translation needed)
+    _RESOURCE_FILTER_FIELDS = {
+        f"{field}__{lookup}"
+        for field, lookups in LatestResourceFilter.Meta.fields.items()
+        for lookup in lookups
+    } | set(LatestResourceFilter.Meta.fields.keys())
+
     def _split_resource_filters(self, params: QueryDict) -> tuple[QueryDict, QueryDict]:
-        resource_keys = set(self.RESOURCE_FILTER_MAP)
+        resource_keys = set(self.RESOURCE_FILTER_MAP) | self._RESOURCE_FILTER_FIELDS
         finding_params = QueryDict(mutable=True)
         resource_params = QueryDict(mutable=True)
         for key, values in params.lists():
@@ -6900,11 +6907,16 @@ def _resource_ids_from_params(
             queryset = queryset.filter(tenant_id=tenant_id)
 
         filter_params = QueryDict(mutable=True)
-        for key, mapped_key in self.RESOURCE_FILTER_MAP.items():
-            if key not in params:
+        for key, values in params.lists():
+            # Translate resource_* prefixed keys via the map
+            if key in self.RESOURCE_FILTER_MAP:
+                mapped_key = self.RESOURCE_FILTER_MAP[key]
+            elif key in self._RESOURCE_FILTER_FIELDS:
+                mapped_key = key
+            else:
                 continue
+
             if key == "resources" or key.endswith("__in"):
-                values = params.getlist(key)
                 items: list[str] = []
                 for value in values:
                     if value is None:
@@ -7232,11 +7244,13 @@ def _build_resource_aggregation(
                 ),
                 first_seen_at=Min("finding__first_seen_at"),
                 last_seen_at=Max("finding__inserted_at"),
-                # Max() picks an arbitrary reason when a resource has multiple
-                # muted findings; this is acceptable because mute rules are
-                # applied per-check so all findings for the same resource
-                # share the same muted_reason in practice.
+                # Max() on muted_reason / check_metadata is safe because
+                # all findings for the same resource+check share identical
+                # values (mute rules and metadata are applied per-check).
                 muted_reason=Max("finding__muted_reason"),
+                resource_group=Max(
+                    KeyTextTransform("resourcegroup", "finding__check_metadata")
+                ),
             )
             .filter(resource_id__isnull=False)
             .order_by("resource_id")
@@ -7273,6 +7287,7 @@ def _post_process_resources(self, resource_data):
                     "first_seen_at": row["first_seen_at"],
                     "last_seen_at": row["last_seen_at"],
                     "muted_reason": row.get("muted_reason"),
+                    "resource_group": row.get("resource_group", ""),
                 }
             )
 
@@ -7307,14 +7322,19 @@ def _build_aggregated_queryset(self, finding_params, latest=False):
             raise ValidationError(filterset.errors)
         filtered_queryset = filterset.qs
         # Only include summaries from each provider's most recent date
-        # (within the filtered range)
-        filtered_queryset = filtered_queryset.annotate(
+        # (within the filtered range).
+        # We use a subquery to strip the Window annotation so it does not
+        # leak into the GROUP BY of _aggregate_daily_summaries.
+        latest_per_provider = filtered_queryset.annotate(
             _max_provider_date=Window(
                 expression=Max("inserted_at"),
                 partition_by=[F("provider_id")],
             ),
         ).filter(inserted_at=F("_max_provider_date"))
-        return self._aggregate_daily_summaries(filtered_queryset)
+        clean_queryset = FindingGroupDailySummary.objects.filter(
+            pk__in=latest_per_provider.values("pk")
+        )
+        return self._aggregate_daily_summaries(clean_queryset)
 
     def _sorted_paginated_response(self, request, aggregated_queryset):
         """Apply ordering, pagination, post-processing, and return the Response."""