fix(apps): always attempt K8s cleanup on delete, cluster service fixes

- Delete endpoint no longer requires clusterId && latestImageTag
- Clusters: getDefault fallback logic, reassign apps on delete
- Unit tests for cluster default/delete logic

backend/src/snapshots/snapshots.service.ts

@@ -351,10 +351,9 @@ export class SnapshotsService {
       }
     }
 
-    // 3. Restore database — file-based
+    // 3. Restore database — file-based (pass file path for PVC-based transfer)
     if (snapshot.dbDumpPath && fs.existsSync(snapshot.dbDumpPath)) {
-      const dumpBuffer = fs.readFileSync(snapshot.dbDumpPath);
-      const result = await this.kubernetesService.restoreDatabaseDump(app, dumpBuffer);
+      const result = await this.kubernetesService.restoreDatabaseDump(app, snapshot.dbDumpPath);
       if (result.success) {
         details.push('✅ Database restored');
         this.logger.log(`Rollback ${snapshotId}: restored database`);