AgentSkillsCN

Kagent Health Analyzer

Kubectl AI 故障排查工具

SKILL.md

kagent Health Analyzer Skill

Purpose

This skill provides cluster-level insights using kagent for analyzing health, optimizing resource allocation, diagnosing pod failures, and generating actionable recommendations based on kagent output.

Capabilities

  • •Execute kagent commands for cluster health analysis
  • •Optimize resource allocation based on cluster usage patterns
  • •Diagnose pod failures and their root causes
  • •Generate comprehensive summaries of kagent output
  • •Provide actionable recommendations for cluster improvements
  • •Interpret kagent analysis results and translate to practical actions

Implementation Details

Cluster Health Analysis

  • •Analyze overall cluster status and component health
  • •Check control plane component statuses (API server, etcd, scheduler, controller manager)
  • •Evaluate node health and resource utilization
  • •Assess networking and storage subsystems
  • •Identify potential bottlenecks or issues

Resource Allocation Optimization

  • •Analyze current resource usage patterns across the cluster
  • •Identify over-provisioned and under-provisioned resources
  • •Recommend resource limit and request adjustments
  • •Suggest node pool scaling based on workload demands
  • •Optimize scheduling efficiency

Pod Failure Diagnostics

  • •Analyze failed pod patterns and common causes
  • •Check for resource contention issues
  • •Identify configuration errors leading to failures
  • •Examine node-related issues affecting pods
  • •Review logs and events for failure patterns

Output Interpretation

  • •Summarize kagent analysis results in a clear format
  • •Highlight critical issues requiring immediate attention
  • •Prioritize recommendations based on severity
  • •Translate technical findings into actionable steps
  • •Provide context for recommendations

Usage

Common kagent Commands:

Cluster Health Analysis:

code
kagent "analyze cluster health"

Resource Optimization:

code
kagent "optimize resource allocation"

Pod Failure Investigation:

code
kagent "why are pods failing"

Comprehensive Analysis:

code
kagent "perform complete cluster analysis"

Analysis Workflow:

Step 1: Execute kagent command

code
kagent_output=$(kagent "{analysis_request}")

Step 2: Parse and summarize results

  • •Extract key metrics and findings
  • •Identify critical issues
  • •Group related problems
  • •Determine severity levels

Step 3: Generate recommendations

  • •Immediate actions for critical issues
  • •Medium-term optimizations
  • •Long-term capacity planning suggestions
  • •Best practice implementations

Example Output Format:

code
Cluster Health Analysis Summary:
- Overall Status: {status}
- Critical Issues Found: {count}
- Resource Utilization: {overview}
- Recommendations: {list}

Detailed Findings:
- {finding_1}: {description}
- {finding_2}: {description}

Recommended Actions:
1. {action_1}: {priority_level}
2. {action_2}: {priority_level}
3. {action_3}: {priority_level}

kagent Command Examples:

For Cluster Health:

code
kagent "analyze cluster health and provide a summary of any issues"

For Resource Optimization:

code
kagent "analyze resource allocation and suggest optimizations for CPU and memory"

For Pod Failures:

code
kagent "investigate recent pod failures and explain the most common causes"

For Capacity Planning:

code
kagent "analyze current usage trends and recommend scaling actions"

Recommended Actions Framework:

Immediate (0-1 hour):

  • •Critical security patches
  • •Failed components requiring restart
  • •Resource exhaustion issues

Short-term (1-24 hours):

  • •Configuration corrections
  • •Minor resource adjustments
  • •Pod restarts/redeployments

Medium-term (1-7 days):

  • •Resource limit optimizations
  • •Node pool adjustments
  • •Storage optimizations

Long-term (1-4 weeks):

  • •Architecture improvements
  • •Scaling strategy implementations
  • •Monitoring enhancements