vllm-project / vllm-project/aibrix

Support Pauseable Upgrades in Storm Service

Open
#1,291 1 comment 0 reactions 0 assignees View on GitHub
area/orchestration help wanted kind/feature priority/important-longterm
Dominant language
Go
Stars
5.1k
Forks
694
Avg merge
1d 19h
Merged PRs (30d)
98

Description

### 🚀 Feature Description and Motivation

Currently, Storm Service lacks controlled upgrade capabilities such as pausing, resuming, or rolling back deployments during the upgrade process. This limitation makes it challenging to implement risk-aware deployment strategies (e.g., canary releases, blue-green deployments) and introduces operational risks when deploying changes to production environments.

This feature request proposes adding upgrade orchestration capabilities similar to [Argo Rollouts](https://argoproj.github.io/argo-rollouts/), including the ability to:

- Pause upgrades at predefined stages for manual verification
- Automatically roll back on failed health checks
- Perform canary releases with traffic routing
- View upgrade progress and status

### Use Case

Support production-grade upgrade experiences

### Proposed Solution

_No response_

Contributor guide

Open the contributing guide

Research direction

The issue does not name any implementation files, tests, or entry points. Start by locating Storm Service's upgrade orchestration, health-check, and traffic-routing paths, then define the scope and acceptance criteria for pausing, resuming, rollback, canary routing, and progress reporting.

Written by the indexing model from the issue text.

Assessment

Tech stack
go
Domain
devops, infrastructure
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.