kthena
https://github.com/volcano-sh/kthena
Go
Lightweight, Modular, Kubernetes-native AI serving platform for scalable model serving.
Triage Issues!
When you volunteer to triage issues, you'll receive an email each day with a link to an open issue that needs help in this project. You'll also receive instructions on how to triage issues.
Triage Docs!
Receive a documented method or class from your favorite GitHub repos in your inbox every day. If you're really pro, receive undocumented methods or classes and supercharge your commit history.
Go not yet supported1 Subscribers
Add a CodeTriage badge to kthena
Help out
- Issues
- router: kvcache GC knobs are unexported, the loop is unstoppable, and errors are silent below v=4
- Support NVIDIA Dynamo workloads with Kthena
- fix(downloader): merge child stderr into stdout to avoid pipe deadlock (#1324)
- docs: document volcano queue placement and node targeting for ModelServing (#899)
- Support lossless upgrade for ModelServing
- model-serving: build the plugin chain once instead of per role replica
- kthena-router exits immediately on SIGTERM: DRAIN_TIMEOUT never applies, in-flight requests are cut and Redis on-flight counts leak
- Add TensorRT-LLM as first-class InferenceEngine (serve, kvcache-aware, PD)
- ModelServing: a persistently failing Role or ServingGroup is recreated endlessly with no backoff
- autoscaler: fix replicas unit mismatch
- Docs
- Go not yet supported