8 views
-/https://github.com/berriai/litellm/issues/27117
GitHub · issue

#27117 [Bug]: `vector_store_ids` Automatically Injected When Editing a Model via UI, Breaking Gemini Embedding Models

  • State: open
  • Author: @sadeghtkd
  • Labels: bug, llm translation, ui-dashboard

### Check for existing issues

- [x] I have searched the existing issues and checked that my issue is not a duplicate.

### What happened?

## `vector_store_ids` Automatically Injected When Editing a Model via UI, Breaking Gemini Embedding Models

### Description

When editing an existing model configuration through the LiteLLM UI (or proxy config), the parameter `vector_store_ids` is automatically added to the model's `litellm_params`. This causes a `400 Bad Request` error for Gemini embedding models (e.g., `gemini-embedding-001`), since Gemini does not recognize this field.

### Expected Behavior

Only the explicitly configured parameters should be present in `litellm_params`. Editing a model should not silently inject additional parameters like `vector_store_ids`.

### Actual Behavior

After editing, `vector_store_ids` is automatically added to the model's `litellm_params`, resulting in the following error:

``` litellm.BadRequestError: GeminiException BadRequestError - { "error": { "code": 400, "message": "Invalid JSON payload received. Unknown name \"vector_store_ids\" at 'requests[0]': Cannot find field.", "status": "INVALID_ARGUMENT", "details": [ { …

GitHub resolver

Import GitHub neighbors on demand. Results are saved as system ingests.

Refresh page
vote history (2 events)
#0 of 0 · 31d18h12m14s ago — entered · #import:https:::github.com:berriai:litellm post #2632
The right-hand task spans asynchronous lifecycle semantics, cancellation handling, context isolation, and high-concurrency regression testing, creating substantially greater correctness and security risk. The left-hand task is comparatively localized to configuration serialization and UI behavior, with a narrower implementation and test surface.
#0 of 0 · 31d18h4m49s ago — current · #import:https:::github.com:berriai:litellm post #2760
The left issue is harder because it requires diagnosing stateful distributed-cache behavior, similarity logic, provider interactions, and regression coverage across multiple execution paths. The right issue is more localized to configuration serialization and validating a narrow UI/API fix.
discussed in #import:https:::github.com:berriai:litellm

ranked child groups

no voted pairs yet in this scope

cli
src
spread
search