maintenance mode + vector store temp db warning

app/api/server.py

@@ -17,6 +17,7 @@ from app.core.vector_store.deps import init_vector_store_db
 from app.services.embeddings_service import EmbeddingService
 from app.services.vector_store_service import VectorStoreService
 from app.api.v1.router import api_v1_router
+from app.api.v1.system import is_maintenance
 
 _logger = get_logger(__name__)
 _settings = get_settings()
@@ -102,6 +103,26 @@ def create_application() -> FastAPI:
         allow_headers=["*"],
     )
 
+    @app.middleware("http")
+    async def maintenance_middleware(request: Request, call_next):
+        if is_maintenance():
+            method = request.method
+            path = request.url.path
+            if method not in ("GET", "HEAD", "OPTIONS"):
+                if not path.startswith("/api/v1/system/maintenance"):
+                    if method == "POST" and path.startswith("/api/v1/backup"):
+                        pass  # allow backup/restore during maintenance
+                    else:
+                        from starlette.responses import JSONResponse
+                        return JSONResponse(
+                            status_code=503,
+                            content={
+                                "success": False,
+                                "detail": "Service is under maintenance. No write operations allowed.",
+                            },
+                        )
+        return await call_next(request)
+
     app.include_router(api_v1_router, prefix="/api/v1")
 
     @app.get("/", include_in_schema=False)

app/api/v1/system.py

@@ -36,6 +36,13 @@ _settings = get_settings()
 _START_TIME = time.time()
 _logger = get_logger(__name__)
 
+_MAINTENANCE_MODE: bool = False
+
+
+
+def is_maintenance() -> bool:
+    return _MAINTENANCE_MODE or os.environ.get("APP_UNDER_MAINTENANCE", "").lower() in ("1", "true", "yes")
+
 
 @router.get("/health", response_model=HealthResponse, summary="Health check")
 async def health(
@@ -185,3 +192,28 @@ async def upload_and_restore(
         "files_restored": restored,
         "data_dir": str(data_dir),
     }
+
+
+@router.post("/maintenance/enable", summary="Enable maintenance mode (blocks write operations)")
+async def enable_maintenance(token: str = Depends(require_auth)):
+    global _MAINTENANCE_MODE
+    _MAINTENANCE_MODE = True
+    _logger.warning("Maintenance mode ENABLED — all write operations blocked")
+    return {"success": True, "message": "Maintenance mode enabled", "maintenance": True}
+
+
+@router.post("/maintenance/disable", summary="Disable maintenance mode")
+async def disable_maintenance(token: str = Depends(require_auth)):
+    global _MAINTENANCE_MODE
+    _MAINTENANCE_MODE = False
+    _logger.warning("Maintenance mode DISABLED — write operations resumed")
+    return {"success": True, "message": "Maintenance mode disabled", "maintenance": False}
+
+
+@router.get("/maintenance", summary="Check maintenance mode status")
+async def maintenance_status(token: str = Depends(require_auth)):
+    return {
+        "success": True,
+        "maintenance": is_maintenance(),
+        "source": "api_flag" if _MAINTENANCE_MODE else ("env_var" if os.environ.get("APP_UNDER_MAINTENANCE") else "off"),
+    }

app/api/v1/vector_stores.py

@@ -101,7 +101,6 @@ async def list_vector_stores(
             document_count=stats.get("document_count", 0),
             created_at=r.created_at,
             metadata=r.metadata,
-            warning=get_temp_db_warning(),
         ))
     return VectorStoreListResponse(success=True, total=len(stores), stores=stores)
 
@@ -130,7 +129,6 @@ async def get_vector_store(
         document_count=stats["document_count"],
         created_at=stats["created_at"],
         metadata=stats["metadata_json"],
-        warning=get_temp_db_warning(),
     )
 
 
@@ -158,7 +156,6 @@ async def delete_vector_store(
         document_count=0,
         embedding_dimension=0,
         created_at=record.created_at,
-        warning=get_temp_db_warning(),
     )
 
 

deploy.py

@@ -256,6 +256,19 @@ def http_get_json(
         return None
 
 
+def http_post_json(
+    url: str, headers: dict[str, str], data: Optional[bytes] = None, timeout: int = 30
+) -> Optional[int]:
+    try:
+        req = urllib.request.Request(url, data=data, headers=headers, method="POST")
+        with urllib.request.urlopen(req, timeout=timeout) as resp:
+            return resp.status
+    except urllib.error.HTTPError as e:
+        return e.code
+    except Exception:
+        return None
+
+
 def http_head(url: str, timeout: int = 30) -> Optional[int]:
     try:
         req = urllib.request.Request(url, method="HEAD")
@@ -562,15 +575,27 @@ def run_deployment(args: argparse.Namespace, env: dict[str, str]) -> int:
         }
 
     # ──────────────────────────────────────────────
-    # PHASE 1: BACKUP (remote via API, fallback to local)
+    # PHASE 1: LOCK WRITES + BACKUP
     # ──────────────────────────────────────────────
-    print(f"\n{_bold('PHASE 1/4: Data Backup')}")
+    print(f"\n{_bold('PHASE 1/4: Lock Writes & Backup')}")
     print(f"  {_INFO_ICON} Backing up data from: {space_url}")
 
     if skip_backup:
         print(f"  {_WARN_ICON} Backup skipped (--skip-backup)")
         audit.set("backup_status", "skipped")
     else:
+        # Enable maintenance mode to freeze data state
+        maint_url = _build_url(space_url, "/api/v1/system/maintenance/enable")
+        maint_enabled = False
+        try:
+            status = http_post_json(maint_url, auth_headers, timeout=15)
+            if status and status == 200:
+                maint_enabled = True
+                print(f"  {_OK_ICON} Maintenance enabled — writes blocked")
+        except Exception:
+            print(f"  {_WARN_ICON} Could not enable maintenance (endpoint may not exist yet)")
+            print(f"  {_WARN_ICON} Data may be lost if writes occur during deployment")
+
         backup_file = persistence_dir / BACKUP_ARCHIVE
         backup_url = _build_url(space_url, BACKUP_ENDPOINT)
         audit.set("backup_started", _timestamp())
@@ -636,7 +661,8 @@ def run_deployment(args: argparse.Namespace, env: dict[str, str]) -> int:
             audit.set("backup_size", total_bytes)
             audit.set("backup_status", "passed")
             print(f"  {_OK_ICON} Files: {file_count} | Size: {_human_size(total_bytes)}")
-            print(f"  {_WARN_ICON} Data added after this point WILL BE LOST until restore completes")
+            if maint_enabled:
+                print(f"  {_OK_ICON} Writes frozen — backup is consistent")
 
     # ──────────────────────────────────────────────
     # PHASE 2: GIT COMMIT & PUSH