gke-ai-troubleshooting-handle-disruption-gpu-tpu
Pass
Audited by Gen Agent Trust Hub on Oct 20, 2026
Risk Level: SAFE
Full Analysis
- Standard Cloud Diagnostics: The skill utilizes standard Kubernetes (
kubectl) commands and Google Cloud Monitoring (PromQL) queries to inspect node state and maintenance labels. These are well-defined diagnostic tools within the GKE ecosystem and follow least-privilege principles. - Workload Protection Guidance: Recommendations for mitigating disruptions (Graceful Termination, Opportunistic Maintenance, and PodDisruptionBudgets) align with official Google Cloud documentation and represent standard operational best practices for maintaining application availability.
- Informative Reporting: The diagnostic workflow emphasizes reporting high-signal information and specifically instructs against dumping raw logs, which helps maintain concise and relevant troubleshooting output.
- Trusted Documentation: All external links point to official
docs.cloud.google.comresources, ensuring that users are directed to authoritative and verified guidance for GKE configuration.
Audit Metadata