ci: Change to use github runner set for ci runners

.github/workflows/nightly-benchmark.yml

@@ -81,7 +81,7 @@ jobs:
     name: "${{ matrix.model.id }} / single-${{ matrix.variant.id }}"
     needs: build-wheel
     if: ${{ !cancelled() }}
-    runs-on: ["k8s-runner-gpu", "4-gpu-h100"]
+    runs-on: 4-gpu-h100
     timeout-minutes: 1440
     strategy:
       fail-fast: false
@@ -188,7 +188,7 @@ jobs:
     name: "${{ matrix.model.id }} / multi-${{ matrix.variant.id }}"
     needs: build-wheel
     if: ${{ !cancelled() }}
-    runs-on: ["k8s-runner-gpu", "4-gpu-h100"]
+    runs-on: 4-gpu-h100
     timeout-minutes: 1440
     strategy:
       fail-fast: false

scripts/k8s-runner-resources/README.md

@@ -0,0 +1,135 @@
+# ARC (Actions Runner Controller) Deployment Guide
+
+## Prerequisites
+
+- Kubernetes cluster (v1.28+)
+- `kubectl` configured with cluster access
+- `helm` v3 installed
+
+## 1. Create a GitHub App
+
+1. Go to **GitHub Organization Settings > Developer settings > GitHub Apps > New GitHub App**
+2. Set the following:
+   - **GitHub App name**: e.g. `smg-k8s-arc-app`
+   - **Homepage URL**: any valid URL
+   - **Webhook**: uncheck "Active" (not needed)
+3. Grant **Repository permissions**:
+   - **Actions**: Read and Write
+   - **Administration**: Read and Write
+   - **Checks**: Read and Write
+   - **Metadata**: Read-only
+4. Click **Create GitHub App**
+5. Note the **App ID** from the app's settings page
+6. Generate a **Private Key** (PEM file) and save it
+
+## 2. Install the GitHub App
+
+1. From the GitHub App settings, click **Install App**
+2. Select the organization and choose the target repository (e.g. `lightseekorg/smg`)
+3. Click **Install**
+4. Note the **Installation ID** from the URL: `https://github.com/organizations/<org>/settings/installations/<installation-id>`
+
+## 3. Create the Kubernetes Secret
+
+Edit `arc-secret-template.yaml` with the values collected above:
+
+```yaml
+apiVersion: v1
+kind: Secret
+metadata:
+  name: github-arc-secret
+  namespace: actions-runner-system
+type: Opaque
+stringData:
+  github_app_id: "<your-app-id>"
+  github_app_installation_id: "<your-installation-id>"
+  github_app_private_key: "<contents-of-your-pem-file>"
+```
+
+Create the namespace and secret:
+
+```bash
+kubectl create namespace actions-runner-system
+kubectl apply -f scripts/k8s-runner-resources/arc-secret-template.yaml
+```
+
+## 4. Install the ARC Controller
+
+```bash
+helm install arc \
+  --namespace actions-runner-system \
+  oci://ghcr.io/actions/actions-runner-controller-charts/gha-runner-scale-set-controller
+```
+
+Verify the controller is running:
+
+```bash
+kubectl get pods -n actions-runner-system -l app.kubernetes.io/name=gha-runner-scale-set-controller
+```
+
+## 5. Install Runner Scale Sets
+
+Install each runner set using its values file:
+
+```bash
+# CPU runners
+helm install k8s-runner-cpu \
+  --namespace actions-runner-system \
+  -f scripts/k8s-runner-resources/runner-values-cpu.yaml \
+  oci://ghcr.io/actions/actions-runner-controller-charts/gha-runner-scale-set
+
+# 4x GPU H100/A10 runners
+helm install k8s-runner-gpu \
+  --namespace actions-runner-system \
+  -f scripts/k8s-runner-resources/runner-values-4-gpu-general.yaml \
+  oci://ghcr.io/actions/actions-runner-controller-charts/gha-runner-scale-set
+
+# 4x GPU H100 runners
+helm install 4-gpu-h100 \
+  --namespace actions-runner-system \
+  -f scripts/k8s-runner-resources/runner-values-4-gpu-h100.yaml \
+  oci://ghcr.io/actions/actions-runner-controller-charts/gha-runner-scale-set
+
+# 2x GPU H100 runners
+helm install 2-gpu-h100 \
+  --namespace actions-runner-system \
+  -f scripts/k8s-runner-resources/runner-values-2-gpu-h100.yaml \
+  oci://ghcr.io/actions/actions-runner-controller-charts/gha-runner-scale-set
+
+# 1x GPU H100 runners
+helm install 1-gpu-h100 \
+  --namespace actions-runner-system \
+  -f scripts/k8s-runner-resources/runner-values-1-gpu-h100.yaml \
+  oci://ghcr.io/actions/actions-runner-controller-charts/gha-runner-scale-set
+```
+
+## 6. Verify
+
+Check that listener pods are running for each runner set:
+
+```bash
+kubectl get pods -n actions-runner-system
+```
+
+Each runner set should have a listener pod in `Running` state. Runner pods will be created on-demand when workflows target the corresponding `runs-on` label (matching `runnerScaleSetName` in each values file).
+
+## Upgrading
+
+To update a runner set (e.g. after changing values):
+
+```bash
+helm upgrade <runner-set-name> \
+  --namespace actions-runner-system \
+  -f scripts/k8s-runner-resources/<runner-set-values-file>.yaml \
+  oci://ghcr.io/actions/actions-runner-controller-charts/gha-runner-scale-set
+```
+
+## Uninstalling
+
+```bash
+# Remove a runner set
+helm uninstall <runner-set-name> -n actions-runner-system
+
+# Remove the controller (after all runner sets are removed)
+helm uninstall arc -n actions-runner-system
+```