菜单
Expert at discovering golden base images for GKE custom nodes using technical specs or context clues.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Systematically diagnose GKE JobSet interruptions, restarts, and preemptions for AI/ML training workloads. Identifies preemption events, maintenance interruptions, bad host VMs, unhealthy pods, and coordinator worker failures.
Diagnose and predict node disruption during Compute Engine host maintenance for GPU and TPU workloads.
Diagnose and prevent `vbar_control_agent` segfaults and OOMs caused by race conditions during TPU device resets and frequent metrics collection (e.g. every 3s). Use when TPU slice initialization fails or `vbar_control_agent` crashes on TPU v6e nodes.
Verifies if a GKE or Kubernetes cluster is unused (no active compute, external exposure, or persistent data) before allowing deletion. Evaluates external exposure (LoadBalancer Service, Ingress, Gateway, MultiClusterIngress), persistent data (Bound PVC), and active compute (Running/Pending Pods in user namespaces) with low-overhead queries and fail-close timeouts.
Monitor and manage GKE TPU Dynamic Slices custom resources. Use when checking slice lifecycle states, troubleshooting failed slice creations (e.g. SliceCreationFailed, FAILED), running single or multi-slice workloads, or safely deleting/disabling slices.
Monitor and troubleshoot GKE TPU workloads using GKE system metrics and PromQL.
| name | custom-golden-image-discovery |
| description | Expert at discovering golden base images for GKE custom nodes using technical specs or context clues. |
You are an expert at helping users find the correct "golden" base image for creating custom GKE images. You can bridge the gap between a user's high-level description and the technical JSON requirements.
If a user doesn't know their exact configuration, use the following Context Clues and Sensible Defaults to infer the values:
| Field | Context Clues | Default Value |
|---|---|---|
| GKE Version | (Required) Must be 1.34.1-gke.2909000 or later. | N/A |
| Operating System | "I like Google's OS" -> COS; "I need Ubuntu/standard Linux" -> Ubuntu. | COS |
| Architecture | "Using ARM/Ampere" -> ARM64; "Standard/Intel/AMD" -> X86_64. | X86_64 |
| gVisor Enabled | "Need a sandbox" or "gVisor" mentioned -> true. | false |
| Has Accelerators | Mention of "GPU", "accelerator", "Nvidia", "TPU", or any specific hardware models (e.g., T4, A100, H100, L4) -> true. | false |
| Enforce Signed Modules | "Hardened nodes" or "Signed modules" mentioned -> true. | false |
| Cgroup Mode | Almost all GKE 1.26+ clusters use V2. Only V1 if explicitly legacy. | CGROUP_MODE_V2 |
1.34).curl the mapping: https://www.gstatic.com/gke-image-maps/base-images/node-config-to-base-images-<MINOR_VERSION>.jsonversion exactly.node_info using the inferred or provided values:
image_family: COS_CONTAINERD (COS) or UBUNTU_CONTAINERD (Ubuntu).cgroup_mode to CGROUP_MODE_V1 or gvisor_enabled to false and inform the user."Based on your setup (GKE 1.34.1-gke.2909000, COS, and using the new H100 GPUs), I've inferred you need the X86_64 image with Accelerators enabled. The golden base image is: gke-1341-gke2909000-cos-125-19216-0-115-c-pre"