fix(proxy/batches): fail closed with 503 when the managed-file lookup errors

A lookup exception previously fell back to dispatching the unresolved
unified token, which defeats the fail-closed guarantee: the token still
reaches the provider and can hit the same publishers-segment IndexError
the resolution prevents. Treat a lookup error like the missing-row case
and fail closed, but with a retryable 503 since the condition is
transient. No-database and no-storage_url rows still fall back unchanged

litellm/proxy/batches_endpoints/endpoints.py

@@ -49,11 +49,12 @@ async def _resolve_managed_input_file_storage_url(input_file_id: str) -> "str |
 
     Provider batch handlers (e.g. Vertex AI, which parses a `publishers/`
     segment out of the file URI) need a real storage location; the opaque
-    unified token crashes them. Returns None when there is no database, the
-    lookup fails, or the row has no storage_url yet, so callers fall back to
-    the original id (which the managed-files deployment hook can still map).
-    Raises a 404 when no managed-file row exists, since the token cannot be
-    resolved and dispatching it would crash the provider.
+    unified token crashes them. Returns None only when there is no database or
+    the row has no storage_url yet, so callers fall back to the original id
+    (which the managed-files deployment hook can still map). Fails closed
+    rather than dispatch a token that cannot be resolved: 404 when no
+    managed-file row exists, 503 when the lookup itself errors so the caller
+    can retry.
     """
     from litellm.proxy.proxy_server import prisma_client
 
@@ -63,7 +64,10 @@ async def _resolve_managed_input_file_storage_url(input_file_id: str) -> "str |
         db_file = await ManagedFileRepository(prisma_client).table.find_first(where={"unified_file_id": input_file_id})
     except Exception as e:
         verbose_proxy_logger.warning("create_batch: managed file lookup failed for %s: %s", input_file_id, e)
-        return None
+        raise HTTPException(
+            status_code=503,
+            detail={"error": "Could not resolve managed file; please retry"},
+        )
     if db_file is None:
         raise HTTPException(
             status_code=404,

tests/test_litellm/proxy/batches_endpoints/test_endpoints.py

@@ -561,10 +561,10 @@ async def test_create__unified_file_id_resolves_real_storage_url(harness):
 
 
 @pytest.mark.asyncio
-async def test_create__unified_file_id_db_error_falls_back_to_raw_id(harness):
-    """A lookup failure must not abort batch creation. Resolution is best-effort
-    crash prevention, so on a database error it falls back to dispatching the
-    original id, which the managed-files deployment hook can still map."""
+async def test_create__unified_file_id_db_error_fails_closed_503(harness):
+    """A lookup error leaves the token unresolved, so it fails closed with a
+    retryable 503 rather than dispatching the opaque id into the provider crash
+    it cannot parse. Nothing is dispatched."""
     set_body(
         harness,
         {
@@ -585,9 +585,12 @@ async def test_create__unified_file_id_db_error_falls_back_to_raw_id(harness):
         patch.object(proxy_server, "prisma_client", MagicMock()),
         patch.object(endpoints, "ManagedFileRepository", fake_repo_cls),
     ):
-        await call_create(harness)
+        with pytest.raises(ProxyException) as exc:
+            await call_create(harness)
 
-    assert harness.router_kwargs()["input_file_id"] == "litellm_proxy_unified_id"
+    assert exc.value.code == "503"
+    harness.router_acreate.assert_not_called()
+    harness.litellm_acreate.assert_not_called()
 
 
 @pytest.mark.asyncio