feat: Make NVIDIA device plugin deployment optional via feature gate

[Copilot]

Kaito unconditionally deploys the NVIDIA device plugin DaemonSet, causing duplicate deployments when clusters already have it installed via GPU Operator or other means. This PR makes the deployment optional via a feature gate, allowing users to disable it when needed.

Changes

• Added nvidiaDevicePlugin.enabled flag to values.yaml (default: true)
• Wrapped device plugin DaemonSet template with conditional rendering
• Updated chart README with configuration documentation

Rationale

Defaulting to enabled ensures Kaito works out of the box for users without GPU Operator. Users with existing device plugin installations can explicitly disable it to prevent conflicts.

Usage

# Default: device plugin deployed by Kaito
helm install kaito-workspace kaito/workspace

# For clusters with existing device plugin (e.g., GPU Operator)
helm install kaito-workspace kaito/workspace --set nvidiaDevicePlugin.enabled=false

Original prompt

---

This section details on the original issue you should resolve

<issue_title>kaito deploys nvidia device plugin when cluster already includes it</issue_title>
<issue_description>Describe the bug

kaito deploys nvidia device plugin when cluster already includes it

they are also deployed on the cpu nodes

Steps To Reproduce

helm install kaito-workspace kaito/workspace --namespace kaito-workspace --create-namespace

Expected behavior

Logs

Environment

• Kubernetes version (use kubectl version):
• OS (e.g: cat /etc/os-release):
• Install tools:
• Others:

Additional context</issue_description>

Comments on the Issue (you are @copilot in this section)

• Fixes kaito deploys nvidia device plugin when cluster already includes it #1700

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.

[kaito-pr-agent]

Title

Make NVIDIA device plugin deployment optional via feature gate

---

Description

• Added feature flag to control NVIDIA device plugin deployment

• Updated Helm chart to conditionally render device plugin DaemonSet

• Documented new configuration option in README

• Set default plugin deployment to enabled for backward compatibility

---

Changes walkthrough 📝

	Relevant files
Documentation	README.md Document NVIDIA device plugin toggle feature                          charts/kaito/workspace/README.md Added documentation for new nvidiaDevicePlugin.enabled configuration parameter Specified default value (true) and usage guidance +1/-0
Configuration changes	nvidia-device-plugin-ds.yaml Add conditional rendering for NVIDIA device plugin              charts/kaito/workspace/templates/nvidia-device-plugin-ds.yaml Wrapped DaemonSet definition in conditional check for nvidiaDevicePlugin.enabled flag Added template directives to enable/disable device plugin deployment +2/-0      values.yaml Implement feature flag for NVIDIA device plugin                    charts/kaito/workspace/values.yaml Added nvidiaDevicePlugin.enabled configuration option Set default value to true for backward compatibility +2/-0

---

Need help?

Type /help how to ... in the comments thread for any questions about PR-Agent usage.

Check out the documentation for more information.

[kaito-pr-agent]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

🎫 Ticket compliance analysis 🔶 1700 - Partially compliant Compliant requirements: Make NVIDIA device plugin deployment optional Provide configuration option to disable plugin deployment Non-compliant requirements: Prevent device plugin from being deployed on CPU nodes Requires further human verification: Verify duplicate deployments are avoided when plugin is disabled Test behavior on clusters with existing NVIDIA device plugin

⏱️ Estimated effort to review: 1 🔵⚪⚪⚪⚪

🧪 No relevant tests

🔒 No security concerns identified

⚡ Recommended focus areas for review Node Selector Missing The DaemonSet lacks node selector to restrict deployment to GPU nodes only metadata: name: nvidia-device-plugin-daemonset namespace: {{ .Release.Namespace }} labels: {{- include "kaito.labels" . | nindent 4 }} spec: selector: matchLabels: name: nvidia-device-plugin-ds updateStrategy: type: RollingUpdate template: metadata: labels: name: nvidia-device-plugin-ds spec: nodeSelector: kubernetes.io/os: linux affinity: nodeAffinity: requiredDuringSchedulingIgnoredDuringExecution: nodeSelectorTerms: - matchExpressions: {{- if eq .Values.cloudProviderName "azure" }} - key: kubernetes.azure.com/cluster operator: Exists - key: type operator: NotIn values: - virtual-kubelet {{- else if eq .Values.cloudProviderName "aws" }} - key: "k8s.io/cloud-provider-aws" operator: Exists {{- end }} tolerations: # Allow this pod to be rescheduled while the node is in "critical add-ons only" mode. # This, along with the annotation above marks this pod as a critical add-on. - key: CriticalAddonsOnly operator: Exists - key: nvidia.com/gpu operator: Exists effect: NoSchedule - key: "sku" operator: "Equal" value: "gpu" effect: "NoSchedule" priorityClassName: "system-node-critical" containers: - image: mcr.microsoft.com/oss/v2/nvidia/k8s-device-plugin:v0.17.0 name: nvidia-device-plugin-ctr env: - name: FAIL_ON_INIT_ERROR value: "false" - name: PASS_DEVICE_SPECS value: "true" securityContext: allowPrivilegeEscalation: false capabilities: drop: ["ALL"] volumeMounts: - name: device-plugin mountPath: /var/lib/kubelet/device-plugins volumes: - name: device-plugin hostPath: path: /var/lib/kubelet/device-plugins

[Copilot]

Pull request overview

This PR addresses the issue where Kaito unconditionally deploys the NVIDIA device plugin DaemonSet, causing conflicts when clusters already have it installed via GPU Operator or other means. The solution introduces a feature gate to make the deployment optional while maintaining backward compatibility by defaulting to enabled.

Key Changes:

• Added nvidiaDevicePlugin.enabled configuration flag (defaulting to true for backward compatibility)
• Wrapped the NVIDIA device plugin DaemonSet template with conditional rendering based on the flag
• Updated the Helm chart README with clear documentation about when to disable the feature

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated no comments.

File	Description
charts/kaito/workspace/values.yaml	Adds the nvidiaDevicePlugin.enabled flag with default value of true to maintain backward compatibility
charts/kaito/workspace/templates/nvidia-device-plugin-ds.yaml	Wraps the entire DaemonSet resource with a conditional check on the new flag
charts/kaito/workspace/README.md	Documents the new configuration option with usage guidance for users with existing device plugin installations

---

The implementation is well-structured and follows existing patterns in the codebase. The conditional rendering syntax matches other feature gates in the repository, the configuration structure is consistent with other top-level settings in values.yaml, and the documentation clearly explains when users should disable this feature. No issues were identified during the review.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

@@           Coverage Diff           @@
##             main    #1707   +/-   ##
=======================================
  Coverage   59.41%   59.41%           
=======================================
  Files          92       92           
  Lines        8697     8697           
=======================================
  Hits         5167     5167           
  Misses       3270     3270           
  Partials      260      260           

Components	Coverage Δ
workspace	49.52% <ø> (ø)
presets	87.10% <ø> (ø)
main	∅ <ø> (∅)

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.