feat(scan, schedule): add next_scan_at field to scans and POST /schedules/daily (#5978)

api/src/backend/tasks/beat.py

@@ -1,7 +1,8 @@
 import json
 
-from django.utils import timezone
-from django_celery_beat.models import PeriodicTask, IntervalSchedule
+from django_celery_beat.models import IntervalSchedule, PeriodicTask
+from rest_framework_json_api.serializers import ValidationError
+from tasks.tasks import perform_scheduled_scan_task
 
 from api.models import Provider
 
@@ -16,7 +17,7 @@ def schedule_provider_scan(provider_instance: Provider):
     task_name = f"scan-perform-scheduled-{provider_instance.id}"
 
     # Schedule the task
-    PeriodicTask.objects.create(
+    _, created = PeriodicTask.objects.get_or_create(
         interval=schedule,
         name=task_name,
         task="scan-perform-scheduled",
@@ -26,6 +27,23 @@ def schedule_provider_scan(provider_instance: Provider):
                 "provider_id": str(provider_instance.id),
             }
         ),
-        start_time=provider_instance.inserted_at + timezone.timedelta(hours=24),
         one_off=False,
     )
+    if not created:
+        raise ValidationError(
+            [
+                {
+                    "detail": "There is already a scheduled scan for this provider.",
+                    "status": 400,
+                    "source": {"pointer": "/data/attributes/provider_id"},
+                    "code": "invalid",
+                }
+            ]
+        )
+
+    return perform_scheduled_scan_task.apply_async(
+        kwargs={
+            "tenant_id": str(provider_instance.tenant_id),
+            "provider_id": str(provider_instance.id),
+        },
+    )

api/src/backend/tasks/tasks.py

@@ -1,5 +1,8 @@
+from datetime import datetime, timedelta, timezone
+
 from celery import shared_task
 from config.celery import RLSTask
+from django_celery_beat.models import PeriodicTask
 from tasks.jobs.connection import check_provider_connection
 from tasks.jobs.deletion import delete_provider
 from tasks.jobs.scan import aggregate_findings, perform_prowler_scan
@@ -98,20 +101,34 @@ def perform_scheduled_scan_task(self, tenant_id: str, provider_id: str):
 
     with tenant_transaction(tenant_id):
         provider_instance = Provider.objects.get(pk=provider_id)
+        periodic_task_instance = PeriodicTask.objects.get(
+            name=f"scan-perform-scheduled-{provider_id}"
+        )
+        next_scan_date = datetime.combine(
+            datetime.now(timezone.utc), periodic_task_instance.date_changed.time()
+        ) + timedelta(hours=24)
 
         scan_instance = Scan.objects.create(
             tenant_id=tenant_id,
             name="Daily scheduled scan",
             provider=provider_instance,
             trigger=Scan.TriggerChoices.SCHEDULED,
+            next_scan_at=next_scan_date,
             task_id=task_id,
         )
 
-    return perform_prowler_scan(
+    result = perform_prowler_scan(
         tenant_id=tenant_id,
         scan_id=str(scan_instance.id),
         provider_id=provider_id,
     )
+    perform_scan_summary_task.apply_async(
+        kwargs={
+            "tenant_id": tenant_id,
+            "scan_id": str(scan_instance.id),
+        }
+    ),
+    return result
 
 
 @shared_task(name="scan-summary")

api/src/backend/tasks/tests/test_beat.py

@@ -0,0 +1,53 @@
+import json
+from unittest.mock import patch
+
+import pytest
+from django_celery_beat.models import IntervalSchedule, PeriodicTask
+from rest_framework_json_api.serializers import ValidationError
+from tasks.beat import schedule_provider_scan
+
+
+@pytest.mark.django_db
+class TestScheduleProviderScan:
+    def test_schedule_provider_scan_success(self, providers_fixture):
+        provider_instance, *_ = providers_fixture
+
+        with patch(
+            "tasks.tasks.perform_scheduled_scan_task.apply_async"
+        ) as mock_apply_async:
+            result = schedule_provider_scan(provider_instance)
+
+            assert result is not None
+
+            mock_apply_async.assert_called_once_with(
+                kwargs={
+                    "tenant_id": str(provider_instance.tenant_id),
+                    "provider_id": str(provider_instance.id),
+                },
+            )
+
+            task_name = f"scan-perform-scheduled-{provider_instance.id}"
+            periodic_task = PeriodicTask.objects.get(name=task_name)
+            assert periodic_task is not None
+            assert periodic_task.interval.every == 24
+            assert periodic_task.interval.period == IntervalSchedule.HOURS
+            assert periodic_task.task == "scan-perform-scheduled"
+            assert json.loads(periodic_task.kwargs) == {
+                "tenant_id": str(provider_instance.tenant_id),
+                "provider_id": str(provider_instance.id),
+            }
+
+    def test_schedule_provider_scan_already_exists(self, providers_fixture):
+        provider_instance, *_ = providers_fixture
+
+        # First, schedule the scan
+        with patch("tasks.tasks.perform_scheduled_scan_task.apply_async"):
+            schedule_provider_scan(provider_instance)
+
+        # Now, try scheduling again, should raise ValidationError
+        with pytest.raises(ValidationError) as exc_info:
+            schedule_provider_scan(provider_instance)
+
+        assert "There is already a scheduled scan for this provider." in str(
+            exc_info.value
+        )