huawei-cloud-cce-autoscaling-diagnoser

What problem does it solve? When autoscaling stops working in a Huawei Cloud CCE cluster, engineers must manually correlate HPA status, Cluster Autoscaler logs, Pending Pod events, node pool limits, subnet IPs, quotas, and IAM permissions across many APIs. This Skill automates that investigation and produces a complete diagnosis report with root cause, confidence, and recommendations. ## Core Features & Use Cases - One-shot autoscaling diagnosis: The huawei_autoscaling_diagnose action performs intent recognition, HPA/CA capability discovery, Path A/B/C routing, and evidence collection, returning a ready-to-use Markdown report. - HPA-to-CA cascade analysis: Links workload-level scaling failures (missing resource requests, maxReplicas reached, metric pipeline broken) with node-level failures (max_nodes reached, subnet IP exhaustion, quota exceeded, IAM denied). - CA Pod log signal matching: Automatically discovers Cluster Autoscaler Pods in kube-system and matches logs against 16 diagnostic signal patterns such as NoExpansionOptions and QuotaExceeded. - Use Case: A user asks why their Deployment is not scaling despite high CPU. The Skill checks the HPA object, finds containers missing CPU resource requests, and reports this as the critical root cause with a remediation suggestion. ## Quick Start Ask the assistant to diagnose why HPA is not scaling your workload in your CCE cluster by providing the region and cluster ID.