An AI-driven diagnostic utility designed to troubleshoot Kubernetes Pod failures, log issues, event storms, and resource exhaustion inside Huawei Cloud CCE clusters.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install huawei-cloud-cce-pod-failure-diagnoser
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install huawei-cloud-cce-pod-failure-diagnoser using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Huawei Cloud CCE Pod Failure Diagnoser?
The Huawei Cloud CCE Pod Failure Diagnoser is a specialized utility designed to identify, analyze, and troubleshoot single-resource Pod failures within Huawei Cloud Cloud Container Engine (CCE) clusters. Running on top of a highly optimized Python dispatcher, this tool bridges the gap between raw cluster metrics and actionable AI insights. It inspects problematic Pod states—including CrashLoopBackOff, ImagePullBackOff, OOMKilled, Pending, and Evicted—without risking cluster stability, functioning as a robust read-only automated investigator.
Integrating this tool with Openclaw Skills empowers developers and intelligent agents to construct logical evidence chains seamlessly. By compiling raw Kubernetes Pod details, system events, container log history, and Cloud Eye/AOM telemetry into structured data, it guarantees that root causes are identified quickly. It offers developers clear, step-by-step remediation suggestions while strictly upholding cloud security configurations and credential safety.
Huawei Cloud CCE Pod Failure Diagnoser Use Cases
- Debugging application workloads stuck in sudden or recurring CrashLoopBackOff states.
- Pinpointing why Kubernetes Pods remain in Pending phases due to failed scheduling or persistent volume claims.
- Investigating ImagePullBackOff or ErrImagePull events to trace image registry authorization and routing conflicts.
- Analyzing OOMKilled container restarts by correlating system memory spikes and thresholds.
- Diagnosing unexpected Pod Evictions caused by localized disk pressure or node instability.
- Compiling rapid health reports and listing abnormal workloads across a targeted Kubernetes namespace.
How Huawei Cloud CCE Pod Failure Diagnoser Works
- Triggering: The developer or intelligent agent identifies failure indicators (e.g., "Pod failed" or "CrashLoopBackOff") and invokes the tool.
- Context and Authentication: The underlying script validates credentials securely from local environment variables (
HUAWEI_AK,HUAWEI_SK, andHUAWEI_REGION) without exposing them in commands. - Kubernetes API Queries: The Python dispatcher script interacts directly with the Huawei Cloud Python SDK and the cluster API to inspect active Pod statuses, historical event cycles, and container specifications.
- Log & Telemetry Harvesting: If enabled, the tool retrieves current logs alongside logs from previous container instances (
previous=true), mapping exit codes directly to potential software faults. - Metric Correlation: When required, resource usage (CPU and memory trends) is fetched via Application Operations Management (AOM) to isolate memory leaks and processing limits.
- Diagnosis Compilation: The collected evidence is consolidated into a comprehensive, sanitized JSON payload containing issue classifications, confidence indicators, and actionable remediation plans.
Huawei Cloud CCE Pod Failure Diagnoser Setup
Prerequisites
Ensure Python 3.8+ is installed in your run environment, and install the required dependencies using pip:
pip install huaweicloudsdkcce huaweicloudsdkcore kubernetes
Credential Configuration
Set up your IAM access keys as secure environment variables. Never hardcode these keys in scripts or config files:
export HUAWEI_AK="<your-huawei-access-key>"
export HUAWEI_SK="<your-huawei-secret-key>"
export HUAWEI_REGION="cn-north-4"
Verifying the Openclaw Skills Installation
Run the read-only test actions to ensure proper network connectivity and IAM authorizations are active:
# Test cluster access and list available Pods
python3 scripts/huawei-cloud.py huawei_get_cce_pods region=cn-north-4 cluster_id=<your-cluster-id>
# Query recent Kubernetes event logs
python3 scripts/huawei-cloud.py huawei_get_cce_events region=cn-north-4 cluster_id=<your-cluster-id> limit=10
Huawei Cloud CCE Pod Failure Diagnoser Data Schema & Taxonomy
Key Input Parameters
| Parameter | Required | Description | Default |
|---|---|---|---|
region |
Yes | Huawei Cloud region identifier | HUAWEI_REGION |
cluster_id |
Yes | Target CCE cluster ID | N/A |
namespace |
No | Targets a specific Kubernetes namespace | default |
pod_name |
No\* | Target Pod name to inspect | N/A |
workload_name |
No\* | Target deployment or workload name | N/A |
include_logs |
No | Collect current and previous logs | true |
include_metrics |
No | Retrieve AOM CPU/Memory telemetry | false |
\* Note: To perform an effective targeted diagnostic, provide at least one identifier among pod_name, workload_name, or resource labels.
Expected JSON Output Structure
The diagnostic engine returns a structured JSON payload detailing its findings:
success: Boolean indicating if the diagnostic script ran successfully.summary.diagnosis_status: High-level status flag (e.g.,abnormal,no_known_failure_detected,no_matching_abnormal_pods).pods[].issues[]: Array of recognized system anomalies, including:type: Category of failure (such asCrashLoopBackOff,OOMKilled,PendingScheduling).confidence: Statistical likelihood of the diagnosed root cause (range 0.0 to 1.0).
top_causes: Ranked list of primary failure candidates with embedded log snippets and configurations.recommended_actions: Read-only commands and inspections for standard resolution.
Huawei Cloud CCE Pod Failure Diagnoser Advanced Features
- Automated Log Sanitizer: Automatically searches for and redacts potential cloud credentials, OAuth tokens, and IAM access keys from terminal logs prior to formatting output.
- Historical Log Diagnostics: Specifically retrieves historical logs from terminated container instances (
previous=true), resolving the standard challenge of finding crash roots for fast-restarting Pods. - AOM Telemetry Lookbacks: Features native integration with Huawei Cloud's Application Operations Management (AOM) service to quickly track performance trends and find TopN resource consumption outliers.
- Alarm-Triggered Inspections: Offers specialized diagnostic entry-points designed to execute workload analyses automatically based on triggered Cloud Eye or custom JSON alerts (
huawei_workload_diagnose_by_alarm). - Collaborative Operator Workflows: Purpose-built as a non-mutative, read-only analytical framework that coordinates easily with mutation-oriented Openclaw Skills like the CCE Auto-Remediation Runner.
SKILL.md
Loading
Related Openclaw Skills
Huawei Cloud CCE Storage Failure Diagnoser
An automated diagnostic tool for troubleshooting Huawei Cloud Cloud Container Engine (CCE) storage, Kubernetes PVC/PV states, and CSI drivers safely and read-only.
Huawei Cloud CCE Root Cause Analyzer
An AI-powered diagnostic skill for Huawei Cloud Cloud Container Engine (CCE) that performs cross-domain root cause analysis and generates comprehensive Markdown incident reports.
Huawei Cloud CCE Workload Failure Diagnoser
A read-only diagnostic skill to analyze rollout failures, replica shortages, and probe failures in Huawei Cloud CCE.
WeChat and WeCom macOS Multi-Instance App Cloner
A highly technical, local-only macOS utility skill for creating, patching, and codesigning fully isolated multi-instance clones of WeChat and WeCom.
Huawei Cloud CCE Pressure Test
An automated performance evaluation and pressure testing tool for Huawei Cloud CCE workloads, leveraging a secure Python SDK dispatcher and k6.
You Need a Budget (YNAB) AI Agent Skill
A powerful AI agent skill implementing Jesse Mecham's proven You Need a Budget (YNAB) methodology to help you break the paycheck cycle and get out of debt.
Featured*
