Starstruk Logo MarkStarstruk

litellm_presidio-dev-v1.81.16

BerriAI/litellmlitellm_presidio-dev-v1.81.16Feb 27, 2026by github-actions[bot]

AI Summary

This release adds pagination and search to /key/aliases to prevent OOMs, introduces control for chunk processing time upperbound, and adds guardrails support for realtime WebSocket sessions. It also includes Prometheus metrics fixes, Datadog APM correlation, and a new /public/endpoints endpoint.

Key Highlights

  • Add pagination and search to /key/aliases to prevent OOMs
  • Control for setting upperbound on chunk processing time
  • UI - Paginated Key Alias Select
  • Guardrails support for Gemini/Vertex AI realtime WebSocket sessions
  • Datadog APM correlation with LiteLLM call IDs

New Features

  • Paginated Key Alias Select in UI
  • /public/endpoints for provider endpoint support
  • Control for chunk processing time upperbound
  • Tags in project
  • Prometheus Metrics Team +Inf Budgets fix

Full Release Notes

## What's Changed
* [Fix] /key/aliases: Add pagination and search to prevent OOMs by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22137
* [Feat] Add control for setting upperbound on chunk processing time  by @ishaan-jaff in https://github.com/BerriAI/litellm/pull/22209
* [Feature] UI - Paginated Key Alias Select by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22157
* fix(mypy): suppress attr-defined errors on realtime websocket calls by @jquinter in https://github.com/BerriAI/litellm/pull/22202
* Litellm prometheus multiproc cleanup by @ryan-crabbe in https://github.com/BerriAI/litellm/pull/22221
* [Fix] Virtual Keys pagination displays stale totals when filtering by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22222
* [Feature] Add /public/supported_endpoints endpoint by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22238
* feat(apm): correlate LiteLLM call IDs with Datadog APM spans by @ishaan-jaff in https://github.com/BerriAI/litellm/pull/22219
* fix(realtime): fix guardrails not firing for Gemini/Vertex AI and provider_config realtime WebSocket sessions by @ishaan-jaff in https://github.com/BerriAI/litellm/pull/22168
* [Fix] Prometheus Metrics Team +Inf Budgets by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22243
* Revert "[Feature] Add /public/supported_endpoints endpoint" by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22246
* fix: remove cache eviction close that kills in-use httpx clients by @ryan-crabbe in https://github.com/BerriAI/litellm/pull/22247
* [Fix] Enhance MidStreamFallbackError to preserve original status code and attributes by @dharamendrak in https://github.com/BerriAI/litellm/pull/22225
* [Infra] Add prisma_schema_sync step as prerequisite for e2e UI tests by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22251
* [Infra] Adding agent_id to Delete Keys Table by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22255
* [Feature] Add /public/endpoints for provider endpoint support by @yuneng-jiang in https://github.com/BerriAI/litellm/pull/22248
* Revert "Fix mapping of parallel_tool_calls for bedrock converse" by @Sameerlite in https://github.com/BerriAI/litellm/pull/22265
* feat: add tags in project by @Harshit28j in https://github.com/BerriAI/litellm/pull/22216


**Full Changelog**: https://github.com/BerriAI/litellm/compare/v1.81.16.custm-auth.dev...litellm_presidio-dev-v1.81.16