project-codeflare / appwrapper

AppWrapper controller for Kueue
https://project-codeflare.github.io/appwrapper/
Apache License 2.0
4 stars 4 forks source link

Disable component-level failure detection for Ray #174

Closed dgrove-oss closed 4 days ago

dgrove-oss commented 4 days ago

In KubeRay 1.1, status.state == failed is not a stable terminal state, therefore we cannot treat it as a signal to initiate a resetOrFail operation on the AppWrapper.