10 Pull requests merged by 4 people

• Configure gpu-deployment.yaml to force vLLM v1 with LoRA

  #573 merged Mar 25, 2025

• Configure the vllm deployment with best practices for startup

  #550 merged Mar 25, 2025

• Bump sigs.k8s.io/controller-runtime from 0.20.3 to 0.20.4

  #570 merged Mar 25, 2025

• Bump github.com/onsi/gomega from 1.36.2 to 1.36.3

  #569 merged Mar 25, 2025

• Bump google.golang.org/protobuf from 1.36.5 to 1.36.6

  #568 merged Mar 25, 2025

• Removing unsafe lib by switching to atomic.Pointer

  #567 merged Mar 25, 2025

• Allow partial metric updates

  #561 merged Mar 24, 2025

• Update boilerplate template

  #566 merged Mar 24, 2025

• Swapping out flow image

  #562 merged Mar 24, 2025

• remove controller-runtime dependency from API

  #565 merged Mar 24, 2025

3 Pull requests opened by 3 people

• Add benchmark automation tool

  #563 opened Mar 24, 2025

• Adding printer columns to inference model

  #574 opened Mar 25, 2025

• Docs: Updates getting started guide for kgateway

  #575 opened Mar 25, 2025

3 Issues closed by 2 people

• Handle response body parsing for both streaming and non-streaming cases

  #178 closed Mar 25, 2025

• Remove Controller-Runtime Dependencies from API Types

  #564 closed Mar 24, 2025

• Refactor the vllm specific code to become model server agnostic

  #383 closed Mar 23, 2025

1 Issue opened by 1 person

• Record the limitation that only a single EPP replica has been tested so far

  #572 opened Mar 25, 2025

9 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• [WIP] Groundwork to support OpenAI API endpoints that vLLM supports

  #526 commented on Mar 24, 2025 • 2 new comments

• update benchmarking guide with latest results with vllm v1

  #559 commented on Mar 24, 2025 • 2 new comments

• Prefix Cache Aware Proposal

  #498 commented on Mar 24, 2025 • 0 new comments

• Expose baseline algorithm parameters as configurable

  #16 commented on Mar 24, 2025 • 0 new comments

• Improve vLLM upstream health checks to only pass when models are servable

  #558 commented on Mar 24, 2025 • 0 new comments

• We should encourage all InferencePool deployments to gracefully rollout and drain

  #549 commented on Mar 25, 2025 • 0 new comments

• Create env vars for the algorithm's scheduling parameters

  #447 commented on Mar 25, 2025 • 0 new comments

• v0.3.0 Release Tracker

  #493 commented on Mar 25, 2025 • 0 new comments

• Validate model/adapter is available on the model server before sending requests to a model server

  #49 commented on Mar 25, 2025 • 0 new comments