vllm-project / vllm-project/aibrix
[ModelClaim] Reschedule claims after terminal runtime failure
Open
Nobody has claimed this yet.
- Dominant language
- Go
- Stars
- 5.1k
- Forks
- 697
- Avg merge
- 1d 19h
- Merged PRs (30d)
- 104
Description
Summary
When an engine exhausts its local restart budget, reschedule the affected ModelClaim to another eligible warm pool Pod and restore routing after activation.
Scope
- Trigger only after the runtime reports a terminal engine failure.
- Keep co-resident engines on the original Pod unaffected.
- Reuse normal placement and activation; no live migration or in-flight request preservation.
- Allow gateway retries to succeed after the new route becomes ready.
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start at the ModelClaim handling and the runtime path that reports a terminal engine failure. Trace normal placement, warm pool Pod activation, and routing restoration; done means only the affected claim moves, co-resident engines remain in place, and gateway retries succeed once the new route is ready.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go
- Domain
- backend, infrastructure
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Quiet
- Clarity
- Mostly clear
- Newbie friendliness
- 45/100