skills.rest
pancake0001
pancake0001COMMUNITY
Explore all skills by this author →
GitHub

huawei-cloud-cce-cost-optimization-advisor

Analyzes Huawei Cloud CCE clusters for idle resources, oversized requests, and HPA optimization opportunities.

1|Updated May 27, 2026
Search Tags:#kubernetes#cost optimization#autoscaler#hpa#huawei cloud#resource utilization#cce
One-click install
npx skills add https://github.com/pancake0001/huawei-cloudnative-skills --skill huawei-cloud-cce-cost-optimization-advisor-pancake0001
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: huawei-cloud-cce-cost-optimization-advisor
Source: https://github.com/pancake0001/huawei-cloudnative-skills/tree/main/releases/container/cce-unscheduled/huawei-cloud-cce-cost-optimization-advisor
Command: npx skills add https://github.com/pancake0001/huawei-cloudnative-skills --skill huawei-cloud-cce-cost-optimization-advisor-pancake0001

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires huaweicloudsdkcce, huaweicloudsdkcore, huaweicloudsdkces, huaweicloudsdkaom, kubernetes, matplotlib, numpy, requests, pyyaml, and includes scripts (resource) and references (resource) components.

What problem does it solve? Kubernetes clusters on Huawei Cloud CCE often waste money through idle nodes, oversized CPU/memory requests, and missing autoscaling policies. This Skill performs read-only cost optimization analysis across 24-hour and 7-day utilization windows and produces actionable recommendations with rollback strategies. ## Core Features & Use Cases - Combined Cost Analysis: Run huawei_analyze_cce_cost_optimization to collect node, pod, and deployment inventory plus AOM Prometheus metrics in one pass, detecting low-utilization nodes and oversized requests with priority classification (high/optimize/observe). - Elasticity Policy Review: Inspect node pool autoscaler settings and existing HPAs, then generate autoscaling/v2 HPA YAML manifests or preview HPA configuration changes. - Safe Write Operations: huawei_configure_cce_hpa returns a preview by default and only applies changes after explicit user confirmation with confirm=true. - Use Case: A platform engineer suspects a CCE cluster is over-provisioned. The Skill analyzes 7-day CPU/memory trends, flags nodes below 30% utilization, lists workloads whose requests exceed 3x actual p95 usage, and outputs a cost optimization report with HPA recommendations. ## Quick Start Ask the assistant to run a cost optimization analysis on your CCE cluster by providing the region and cluster ID, for example: analyze cost optimization opportunities for cluster xxx in cn-north-4 excluding kube-system.

Frequently Asked Questions about huawei-cloud-cce-cost-optimization-advisor

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I analyze CCE cluster cost optimization on Huawei Cloud?▼

Run the combined action huawei_analyze_cce_cost_optimization with your region and cluster_id. It collects node, pod, and deployment inventory plus AOM metrics, then outputs low-utilization nodes, oversized requests, and HPA recommendations in a single report.

How to detect oversized CPU and memory requests in Kubernetes workloads?▼

The Skill compares each business workload's request against actual p95 usage over 24-hour and 7-day windows. Requests exceeding 3x p95 are marked high priority, 2x are marked optimize, and short-window-only matches are marked observe.

Does this tool modify HPA or node pools automatically?▼

No. All analysis is read-only by default. huawei_configure_cce_hpa without confirm=true returns only a manifest preview, and applying changes requires explicit user confirmation with confirm=true. Node pool scale-down and request modification are never performed automatically.

What IAM permissions are needed for CCE cost analysis?▼

You need CCE read permissions such as cce:cluster:list, cce:node:list, and cce:nodepool:list, plus aom:*:get for reading AOM Prometheus metrics. Credentials are supplied via HW_ACCESS_KEY and HW_SECRET_KEY environment variables.

Why does the cost report show data gaps or missing metrics?▼

Data gaps appear when AOM metrics are unavailable, workload requests are absent, or HPA status is invisible. Verify the IAM policy includes aom:*:get, AOM is enabled on the cluster, and the kubernetes Python package is installed for HPA queries.

Can I scale down nodes based on 24-hour low utilization alone?▼

No. The Skill requires both 24-hour and 7-day windows to confirm low utilization before recommending scale-down. A single short-window dip is flagged as short-term fluctuation and marked for observation only.

Related Skills

Explore More Software Engineering Skills→
gke-labsgke-labs
42

gke-observability

Audit and configure logging, monitoring, and tracing for Google Kubernetes Engine clusters.

Official
Intermediate
googlegoogle
17.1k

gke-scaling

Automate GKE workload scaling with HPA, VPA, and Node Auto Provisioning.

Official
Advanced
ohsonerdyohsonerdy

capacity-planning

Calculate fleet sizing with headroom from demand forecasts and load tests.

Community
Advanced
Cure-Consulting-GroupCure-Consulting-Group

finops

Manage FinOps processes for Firebase and GCP projects with budgeting and cost allocation.

Official
Advanced