What problem does it solve? When a Deployment, StatefulSet, or DaemonSet in a Huawei Cloud CCE cluster gets stuck during rollout or loses available replicas, engineers must manually correlate workload status, ReplicaSets, Pods, Events, and logs to find the root cause. This Skill automates that read-only evidence collection and ranks the most likely causes with a structured rollout funnel. ## Core Features & Use Cases - Rollout Funnel Analysis: Walks the chain from control-plane generation observation through new ReplicaSet creation, Pod scheduling, readiness, and availability to pinpoint the first failing layer. - Evidence-Based Cause Ranking: Maps symptoms like ProgressDeadlineExceeded, ImagePullBackOff, CrashLoopBackOff, OOMKilled, FailedScheduling, probe failures, and FailedMount to labeled causes with confidence scores and cited evidence. - Safe Read-Only Diagnosis: Uses hcloud CCE commands for cluster discovery and the kubectl-cce plugin for Kubernetes reads, never mutating resources, and hands off remediation to dedicated skills. - Use Case: A Deployment update stalls with new Pods stuck in Pending. The Skill verifies cluster access, inspects the Deployment and its ReplicaSets, filters Events to the workload, identifies FailedScheduling due to node taints, and recommends handing off to the node failure diagnoser. ## Quick Start Ask the AI to diagnose why Deployment 'api-server' in namespace 'prod' of CCE cluster <cluster-id> in region cn-north-4 is not rolling out, and provide a ranked cause report.
