Skip to content

[POC]: feat: add Kueue prebuilt workload backend#707

Open
alaypatel07 wants to merge 1 commit into
ai-dynamo:mainfrom
alaypatel07:poc/kueue-prebuilt-workload
Open

[POC]: feat: add Kueue prebuilt workload backend#707
alaypatel07 wants to merge 1 commit into
ai-dynamo:mainfrom
alaypatel07:poc/kueue-prebuilt-workload

Conversation

@alaypatel07

Copy link
Copy Markdown

Add a Kueue scheduler backend path that creates prebuilt Workloads from Grove PodGangs so PodGroup minReplicas can map to Kueue PodSet minCount. Include focused demo manifests for standalone minCount and PCSG all-or-nothing behavior.

What type of PR is this?

What this PR does / why we need it:

This PR adds Kueue integration. This PR:

  1. adds kueue scheduler backend
  2. the backend pre-creates the workload for the pods
  3. the operator then creates the pod with specific labels that is needed
    for kueue to admit the workload.

Which issue(s) this PR fixes:

Fixes #704

Special notes for your reviewer:

Does this PR introduce a API change?

GREP-704: add kueue integration

Additional documentation e.g., enhancement proposals, usage docs, etc.:

TBD

@copy-pr-bot

copy-pr-bot Bot commented Jul 9, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@shayasoolin

Copy link
Copy Markdown
Contributor

@enoodle would you like to review this? Thanks

Add a Kueue scheduler backend path that creates prebuilt Workloads from Grove PodGangs so PodGroup minReplicas can map to Kueue PodSet minCount. Include focused demo manifests for standalone minCount and PCSG all-or-nothing behavior.
@alaypatel07 alaypatel07 force-pushed the poc/kueue-prebuilt-workload branch from 5f82b70 to be076f2 Compare July 13, 2026 12:12
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

Kueue integration

2 participants