18 views
-/https://github.com/berriai/litellm/issues/23005
GitHub · issue

#23005 [Bug] UI very slow in version 1.82.x

  • State: open
  • Author: @babanetworx

## Bug Description

The LiteLLM UI is extremely slow in version 1.82.x. The interface takes a long time to load and respond to user interactions.

## Steps to Reproduce

1. Install LiteLLM version 1.82.x 2. Access the UI (localhost:4000 or configured URL) 3. Observe slow loading times and sluggish interface

## Expected Behavior

The UI should be responsive and load quickly, similar to earlier versions.

## Actual Behavior

The UI is very slow - pages take a long time to load and interactions are sluggish.

## Environment

- LiteLLM Version: 1.82.x - Installation method: [pip/docker] - OS: [your OS]

## Relevant Issues

This may be related to: - Issue #19921: Performance regression in 1.81.x (UI + API slowness) - Issue #6345: Performance degradation over time

## Additional Context

Please see related issues for more context on performance issues in recent versions.

GitHub resolver

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

Refresh page
vote history (6 events)
#0 of 0 · 31d19h12m28s ago — entered · #import:https:::github.com:berriai:litellm post #1679
Issue 23005 is harder because it requires profiling and isolating a broad regression across UI rendering, API behavior, and possibly database or asset-loading paths, with unclear reproduction boundaries. Issue 30425 is comparatively contained: dependency packaging, Prisma client generation, and startup validation can be addressed through a focused build or installation change.
29818 requires cross-layer protocol compatibility work, including proxy routing, request/response translation, and streaming behavior for a client-specific API path. 23005 is a broader but more localized performance regression investigation, likely addressed through profiling and UI/backend optimization.
The UI performance problem is harder because it requires broad profiling and regression isolation across potentially many frontend and backend paths, while the router issue has a more localized fix with targeted validation and security-focused tests.
The UI performance issue is harder because it requires broad profiling and regression diagnosis across frontend rendering, API behavior, database access, and deployment environments, whereas the mapper issue is a more localized attribute-mapping change with clearer existing integration behavior.
The UI performance regression is harder because it requires broad profiling and isolation across frontend, API, database, and deployment factors, whereas the cache failure is more narrowly scoped to dependency or connection-parameter compatibility.
#0 of 0 · 31d18h11m51s ago — current · #import:https:::github.com:berriai:litellm post #2663
The left requires coordinated changes across asynchronous control flow, failure propagation, protocol-compliant terminal events, and regression coverage. The right is broad but primarily calls for profiling and targeted performance remediation, with less clearly defined implementation scope.
discussed in #import:https:::github.com:berriai:litellm

ranked child groups

no voted pairs yet in this scope

cli
src
spread
search