vllm-project / vllm-project/agentic-api

Enterprise Readiness

Open
#316 0 comments 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

enhancement enterprise-readiness
Dominant language
Rust
Stars
289
Forks
74
Avg merge
1d 17h
Merged PRs (30d)
93

Description

Problem statement / motivation

Enterprise Readiness groups the enhancements needed to operate vLLM Agentic API as a secure, predictable service for Responses and Messages workloads. The focus is authenticated state ownership, reliable execution, protocol compatibility, complete usage accounting, and repeatable release qualification.

Proposed solution

Track the work in the existing focused issues where available, with new issues for capabilities that did not yet have an owner. These are feature enhancements; completion is determined by the linked acceptance criteria and validation evidence.

Enhancement checklist
  • #107 — Authenticated user and tenant ownership for persisted state and continuation.
  • #313 — Bounded, validated Messages streaming and tool scheduling.
  • #314 — Request capability and transport metadata preservation across execution paths.
  • #155 — Standard Conversations API references, creation payloads, and resource operations.
  • #145 — Resumable MCP tool approval policies and enforcement.
  • #315 — Complete multi-round Messages usage accounting.
  • #110 — Deployment, authentication, persistence, and protocol qualification through the actual service.
Delivery order
  1. Establish authenticated ownership and predictable Messages lifecycle/resource behavior; develop the request capability contract alongside them.
  2. Complete standard conversation workflows, approval-dependent tool workflows, and usage reporting for the supported deployment profiles.
  3. Qualify those profiles through deterministic integration tests and documented engine coverage before declaring the initiative complete. Build the qualification harness throughout the work, not only at the end.

Shared multi-tenant operation depends on #107. Approval-dependent tools depend on #145. Metered multi-round Messages workloads depend on #315. Narrower deployment profiles can ship with explicit capability limits while this initiative remains open.

Completion criteria
  • Every enhancement above meets its acceptance criteria, or an explicit scope decision updates this tracker and the roadmap.
  • Supported Responses and Messages capabilities, resource limits, ownership rules, and failure behavior are documented.
  • Repeatable release evidence distinguishes mock integration tests, recorded replay, and live-engine coverage, including unsupported scenarios.
  • Single-tenant and supported shared deployments have documented upgrade and operational guidance.
Additional context

This initiative builds on #102 (hosted deployment), #211 (engine conformance), #241/#244 (streaming boundaries), and #279 (observability). Those workstreams retain their own scope. Existing fixes in #290 and #296 and Dynamo Messages recordings in #292 are inputs to the qualification work.

Additional search providers, file-search expansion, new upstream protocol adapters, and broader tool families remain separate capability work unless a supported enterprise deployment profile explicitly requires them.

Roadmap update: https://github.com/vllm-project/agentic-api/pull/317

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

Start by reviewing the focused issues listed in the enhancement checklist, especially their acceptance criteria and validation evidence. Use the delivery order and completion criteria to determine how ownership, protocol behavior, usage accounting, deployment profiles, and qualification evidence fit together. Done means the listed enhancements are completed or explicitly scoped, with supported capabilities and release evidence documented.

Written by the indexing model from the issue text.

Assessment

Tech stack
rust
Domain
api, backend, documentation, security, testing
Issue type
Feature
Difficulty
5/5
Estimated time
Over a week
Activity status
Stale
Clarity
Needs clarification
Newbie friendliness
20/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.