6 views
-/https://github.com/berriai/litellm/issues/32621
GitHub · issue

#32621 [Bug]: Cache Settings UI — "Semantic" listed as a Redis Type alongside Node/Cluster/Sentinel which are network topologies, causing user confusion

  • State: open
  • Author: @vigneshrajaswarise
  • Labels: bug, ui-dashboard

### Check for existing issues

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

### What happened?

In the LiteLLM Admin Dashboard under Experimental → Cache, there is a "Redis Type" dropdown with 4 options:

Node Cluster Sentinel Semantic ← ❌ This does not belong here Node, Cluster, and Sentinel are Redis network deployment topologies — they describe how your Redis infrastructure is set up.

Semantic is a caching strategy — it describes how cache lookups are performed (using embedding vectors instead of exact key matching).

These are completely different dimensions. Mixing "Semantic" into the Redis Type dropdown is like listing "Search by color" alongside "Car", "Truck", and "Motorcycle" — one of these things is not like the others.

<img width="1635" height="736" alt="Image" src="https://github.com/user-attachments/assets/a08d9794-59f8-4c71-a442-27c4fd778fce" />

### Steps to Reproduce

1. Log in to LiteLLM Admin Dashboard 2. Go to Experimental → Caching 3. Select Cache Settings - Open the Redis Type dropdown 4. Observe: Node, Cluster, Sentinel, Semantic are all listed together 5. Notice that Node, Cluster, Sentinel = network topology; Se…

GitHub resolver

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

Refresh page
vote history (3 events)
#0 of 0 · 31d18h15m36s ago — entered · #import:https:::github.com:berriai:litellm post #2582
Adding support across multiple provider model modalities requires API integration, parameter/schema mapping, compatibility handling, and broader testing; the dashboard issue is comparatively localized to UI categorization and related state/configuration handling.
The left task spans backend integration, external service interaction, model metadata, metric computation, error handling, and propagation across multiple observability adapters. The right task is a localized dashboard taxonomy and form-state correction with comparatively limited regression risk.
#0 of 0 · 31d17h54m58s ago — current · #import:https:::github.com:berriai:litellm post #2940
Changing shared provider-resolution precedence carries broader behavioral and regression risk across integrations, while the other issue is a localized dashboard configuration fix.
discussed in #import:https:::github.com:berriai:litellm

ranked child groups

no voted pairs yet in this scope

cli
src
spread
search