bakery-ia/COMPLETION_CHECKLIST.md

# Completion Checklist - Tenant & User Deletion System

**Current Status:** 75% Complete
**Time to 100%:** ~4 hours implementation + 2 days testing

---

## Phase 1: Complete Remaining Services (1.5 hours)

### POS Service (30 minutes)

- [ ] Create `services/pos/app/services/tenant_deletion_service.py`
  - [ ] Copy template from QUICK_START_REMAINING_SERVICES.md
  - [ ] Import models: POSConfiguration, POSTransaction, POSSession
  - [ ] Implement `get_tenant_data_preview()`
  - [ ] Implement `delete_tenant_data()` with correct order:
    - [ ] 1. POSTransaction
    - [ ] 2. POSSession
    - [ ] 3. POSConfiguration

- [ ] Add endpoints to `services/pos/app/api/{router}.py`
  - [ ] DELETE /tenant/{tenant_id}
  - [ ] GET /tenant/{tenant_id}/deletion-preview

- [ ] Test manually:
  ```bash
  curl -X GET "http://localhost:8000/api/v1/pos/tenant/{id}/deletion-preview"
  curl -X DELETE "http://localhost:8000/api/v1/pos/tenant/{id}"
  ```

### External Service (30 minutes)

- [ ] Create `services/external/app/services/tenant_deletion_service.py`
  - [ ] Copy template
  - [ ] Import models: ExternalDataCache, APIKeyUsage
  - [ ] Implement `get_tenant_data_preview()`
  - [ ] Implement `delete_tenant_data()` with order:
    - [ ] 1. APIKeyUsage
    - [ ] 2. ExternalDataCache

- [ ] Add endpoints to `services/external/app/api/{router}.py`
  - [ ] DELETE /tenant/{tenant_id}
  - [ ] GET /tenant/{tenant_id}/deletion-preview

- [ ] Test manually

### Alert Processor Service (30 minutes)

- [ ] Create `services/alert_processor/app/services/tenant_deletion_service.py`
  - [ ] Copy template
  - [ ] Import models: Alert, AlertRule, AlertHistory
  - [ ] Implement `get_tenant_data_preview()`
  - [ ] Implement `delete_tenant_data()` with order:
    - [ ] 1. AlertHistory
    - [ ] 2. Alert
    - [ ] 3. AlertRule

- [ ] Add endpoints to `services/alert_processor/app/api/{router}.py`
  - [ ] DELETE /tenant/{tenant_id}
  - [ ] GET /tenant/{tenant_id}/deletion-preview

- [ ] Test manually

---

## Phase 2: Refactor Existing Services (2.5 hours)

### Forecasting Service (45 minutes)

- [ ] Review existing deletion logic in forecasting service
- [ ] Create new `services/forecasting/app/services/tenant_deletion_service.py`
  - [ ] Extend BaseTenantDataDeletionService
  - [ ] Move existing logic into standard pattern
  - [ ] Import models: Forecast, PredictionBatch, etc.

- [ ] Update endpoints to use new pattern
  - [ ] Replace existing DELETE logic
  - [ ] Add deletion-preview endpoint

- [ ] Test both endpoints

### Training Service (45 minutes)

- [ ] Review existing deletion logic
- [ ] Create new `services/training/app/services/tenant_deletion_service.py`
  - [ ] Extend BaseTenantDataDeletionService
  - [ ] Move existing logic into standard pattern
  - [ ] Import models: TrainingJob, TrainedModel, ModelArtifact

- [ ] Update endpoints to use new pattern

- [ ] Test both endpoints

### Notification Service (45 minutes)

- [ ] Review existing deletion logic
- [ ] Create new `services/notification/app/services/tenant_deletion_service.py`
  - [ ] Extend BaseTenantDataDeletionService
  - [ ] Move existing logic into standard pattern
  - [ ] Import models: Notification, NotificationPreference, etc.

- [ ] Update endpoints to use new pattern

- [ ] Test both endpoints

---

## Phase 3: Integration (2 hours)

### Update Auth Service

- [ ] Open `services/auth/app/services/admin_delete.py`

- [ ] Import DeletionOrchestrator:
  ```python
  from app.services.deletion_orchestrator import DeletionOrchestrator
  ```

- [ ] Update `_delete_tenant_data()` method:
  ```python
  async def _delete_tenant_data(self, tenant_id: str):
      orchestrator = DeletionOrchestrator(auth_token=self.get_service_token())
      job = await orchestrator.orchestrate_tenant_deletion(
          tenant_id=tenant_id,
          tenant_name=tenant_info.get("name"),
          initiated_by=self.requesting_user_id
      )
      return job.to_dict()
  ```

- [ ] Remove old manual service calls

- [ ] Test complete user deletion flow

### Verify Service URLs

- [ ] Check orchestrator SERVICE_DELETION_ENDPOINTS
- [ ] Update URLs for your environment:
  - [ ] Development: localhost ports
  - [ ] Staging: service names
  - [ ] Production: service names

---

## Phase 4: Testing (2 days)

### Unit Tests (Day 1)

- [ ] Test TenantDataDeletionResult
  ```python
  def test_deletion_result_creation():
      result = TenantDataDeletionResult("tenant-123", "test-service")
      assert result.tenant_id == "tenant-123"
      assert result.success == True
  ```

- [ ] Test BaseTenantDataDeletionService
  ```python
  async def test_safe_delete_handles_errors():
      # Test error handling
  ```

- [ ] Test each service deletion class
  ```python
  async def test_orders_deletion():
      # Create test data
      # Call delete_tenant_data()
      # Verify data deleted
  ```

- [ ] Test DeletionOrchestrator
  ```python
  async def test_orchestrator_parallel_execution():
      # Mock service responses
      # Verify all called
  ```

- [ ] Test DeletionJob tracking
  ```python
  def test_job_status_tracking():
      # Create job
      # Check status transitions
  ```

### Integration Tests (Day 1-2)

- [ ] Test tenant deletion endpoint
  ```python
  async def test_delete_tenant_endpoint():
      response = await client.delete(f"/api/v1/tenants/{tenant_id}")
      assert response.status_code == 200
  ```

- [ ] Test service-to-service calls
  ```python
  async def test_orders_deletion_via_orchestrator():
      # Create tenant with orders
      # Delete tenant
      # Verify orders deleted
  ```

- [ ] Test CASCADE deletes
  ```python
  async def test_cascade_deletes_children():
      # Create parent with children
      # Delete parent
      # Verify children also deleted
  ```

- [ ] Test error handling
  ```python
  async def test_partial_failure_handling():
      # Mock one service failure
      # Verify job shows failure
      # Verify other services succeeded
  ```

### E2E Tests (Day 2)

- [ ] Test complete tenant deletion
  ```python
  async def test_complete_tenant_deletion():
      # Create tenant with data in all services
      # Delete tenant
      # Verify all data deleted
      # Check deletion job status
  ```

- [ ] Test complete user deletion
  ```python
  async def test_user_deletion_with_owned_tenants():
      # Create user with owned tenants
      # Create other admins
      # Delete user
      # Verify ownership transferred
      # Verify user data deleted
  ```

- [ ] Test owner deletion with tenant deletion
  ```python
  async def test_owner_deletion_no_other_admins():
      # Create user with tenant (no other admins)
      # Delete user
      # Verify tenant deleted
      # Verify all cascade deletes
  ```

### Manual Testing (Throughout)

- [ ] Test with small dataset (<100 records)
- [ ] Test with medium dataset (1,000 records)
- [ ] Test with large dataset (10,000+ records)
- [ ] Measure performance
- [ ] Verify database queries are efficient
- [ ] Check logs for errors
- [ ] Verify audit trail

---

## Phase 5: Database Persistence (1 day)

### Create Migration

- [ ] Create deletion_jobs table:
  ```sql
  CREATE TABLE deletion_jobs (
      id UUID PRIMARY KEY,
      tenant_id UUID NOT NULL,
      tenant_name VARCHAR(255),
      initiated_by UUID,
      status VARCHAR(50) NOT NULL,
      service_results JSONB,
      total_items_deleted INTEGER DEFAULT 0,
      started_at TIMESTAMP WITH TIME ZONE,
      completed_at TIMESTAMP WITH TIME ZONE,
      error_log TEXT[],
      created_at TIMESTAMP WITH TIME ZONE DEFAULT NOW()
  );

  CREATE INDEX idx_deletion_jobs_tenant ON deletion_jobs(tenant_id);
  CREATE INDEX idx_deletion_jobs_status ON deletion_jobs(status);
  CREATE INDEX idx_deletion_jobs_initiated ON deletion_jobs(initiated_by);
  ```

- [ ] Run migration in dev
- [ ] Run migration in staging

### Update Orchestrator

- [ ] Add database session to DeletionOrchestrator
- [ ] Save job to database in orchestrate_tenant_deletion()
- [ ] Update job status in database
- [ ] Query jobs from database in get_job_status()
- [ ] Query jobs from database in list_jobs()

### Add Job API Endpoints

- [ ] Create `services/auth/app/api/deletion_jobs.py`
  ```python
  @router.get("/deletion-jobs/{job_id}")
  async def get_job_status(job_id: str):
      # Query from database

  @router.get("/deletion-jobs")
  async def list_deletion_jobs(
      tenant_id: Optional[str] = None,
      status: Optional[str] = None,
      limit: int = 100
  ):
      # Query from database with filters
  ```

- [ ] Test job status endpoints

---

## Phase 6: Production Prep (2 days)

### Performance Testing

- [ ] Create test dataset with 100K records
- [ ] Run deletion and measure time
- [ ] Identify bottlenecks
- [ ] Optimize slow queries
- [ ] Add batch processing if needed
- [ ] Re-test and verify improvement

### Monitoring Setup

- [ ] Add Prometheus metrics:
  ```python
  deletion_duration_seconds = Histogram(...)
  deletion_items_deleted = Counter(...)
  deletion_errors_total = Counter(...)
  deletion_jobs_status = Gauge(...)
  ```

- [ ] Create Grafana dashboard:
  - [ ] Active deletions gauge
  - [ ] Deletion rate graph
  - [ ] Error rate graph
  - [ ] Average duration graph
  - [ ] Items deleted by service

- [ ] Configure alerts:
  - [ ] Alert if deletion >5 minutes
  - [ ] Alert if >10% error rate
  - [ ] Alert if service timeouts

### Documentation Updates

- [ ] Update API documentation
- [ ] Create operations runbook
- [ ] Document rollback procedures
- [ ] Create troubleshooting guide

### Rollout Plan

- [ ] Deploy to dev environment
- [ ] Run full test suite
- [ ] Deploy to staging
- [ ] Run smoke tests
- [ ] Deploy to production with feature flag
- [ ] Monitor for 24 hours
- [ ] Enable for all tenants

---

## Phase 7: Optional Enhancements (Future)

### Soft Delete (2 days)

- [ ] Add deleted_at column to tenants table
- [ ] Implement 30-day retention
- [ ] Add restoration endpoint
- [ ] Add cleanup job for expired deletions
- [ ] Update queries to filter deleted tenants

### Advanced Features (1 week)

- [ ] WebSocket progress updates
- [ ] Email notifications on completion
- [ ] Deletion reports (PDF download)
- [ ] Scheduled deletions
- [ ] Deletion preview aggregation

---

## Sign-Off Checklist

### Code Quality

- [ ] All services implemented
- [ ] All endpoints tested
- [ ] No compiler warnings
- [ ] Code reviewed
- [ ] Documentation complete

### Testing

- [ ] Unit tests passing (>80% coverage)
- [ ] Integration tests passing
- [ ] E2E tests passing
- [ ] Performance tests passing
- [ ] Manual testing complete

### Production Readiness

- [ ] Monitoring configured
- [ ] Alerts configured
- [ ] Logging verified
- [ ] Rollback plan documented
- [ ] Runbook created

### Security & Compliance

- [ ] Authorization verified
- [ ] Audit logging enabled
- [ ] GDPR compliance verified
- [ ] Data retention policy documented
- [ ] Security review completed

---

## Quick Reference

### Files to Create (3 new services):
1. `services/pos/app/services/tenant_deletion_service.py`
2. `services/external/app/services/tenant_deletion_service.py`
3. `services/alert_processor/app/services/tenant_deletion_service.py`

### Files to Modify (3 refactored services):
1. `services/forecasting/app/services/tenant_deletion_service.py`
2. `services/training/app/services/tenant_deletion_service.py`
3. `services/notification/app/services/tenant_deletion_service.py`

### Files to Update (integration):
1. `services/auth/app/services/admin_delete.py`

### Tests to Write (~50 tests):
- 10 unit tests (base classes)
- 24 service-specific tests (2 per service × 12 services)
- 10 integration tests
- 6 E2E tests

### Time Estimate:
- Implementation: 4 hours
- Testing: 2 days
- Deployment: 2 days
- **Total: ~5 days**

---

## Success Criteria

✅ All 12 services have deletion logic
✅ All deletion endpoints working
✅ Orchestrator coordinating successfully
✅ Job tracking persisted to database
✅ All tests passing
✅ Performance acceptable (<5 min for large tenants)
✅ Monitoring in place
✅ Documentation complete
✅ Production deployment successful

---

**Keep this checklist handy and mark items as you complete them!**

**Remember:** Templates and examples are in QUICK_START_REMAINING_SERVICES.md