optimization-monitor（优化监控器）

Performance Monitor Skill

Overview

This skill provides comprehensive real-time performance monitoring capabilities including metrics collection, bottleneck detection, SLA compliance tracking, anomaly detection, and resource utilization monitoring for swarm-based systems.

When to Use

• Continuous monitoring of swarm performance
• Detecting performance bottlenecks before they impact operations
• Tracking SLA compliance and generating alerts
• Anomaly detection in system metrics
• Resource utilization tracking and forecasting
• Building real-time performance dashboards

Quick Start

# Start comprehensive monitoring
npx claude-flow performance-report --format detailed --timeframe 24h

# Real-time bottleneck analysis
npx claude-flow bottleneck-analyze --component swarm-coordination

# Health check all components
npx claude-flow health-check --components ["swarm", "agents", "coordination"]

# Collect specific metrics
npx claude-flow metrics-collect --components ["cpu", "memory", "network"]

Architecture

+-----------------------------------------------------------+
|                  Performance Monitor                       |
+-----------------------------------------------------------+
|  Metrics Collector  |  Bottleneck Analyzer  |  SLA Monitor |
+---------------------+-----------------------+--------------+
         |                     |                      |
         v                     v                      v
+-------------------+  +------------------+  +---------------+
| System Metrics    |  | Pattern Detection|  | Threshold     |
| - CPU/Memory      |  | - CPU Bottleneck |  | Checking      |
| - I/O/Network     |  | - Memory Leak    |  | - Availability|
| - Process Stats   |  | - I/O Saturation |  | - Response    |
+-------------------+  | - Network Issues |  | - Throughput  |
                       +------------------+  +---------------+
         |                     |                      |
         v                     v                      v
+-----------------------------------------------------------+
|              Dashboard Provider (Real-time)                |
+-----------------------------------------------------------+

Core Capabilities

1. Multi-Dimensional Metrics Collection

// Real-time metrics collection
const metrics = await mcp.metrics_collect({
  components: ['cpu', 'memory', 'network', 'agents']
});

// System metrics include:
// - CPU: usage, load average, core utilization
// - Memory: usage, available, pressure
// - I/O: disk usage, disk I/O, network I/O
// - Processes: count, threads, handles

2. Bottleneck Detection

Detects and categorizes bottlenecks:

• CPU Bottlenecks: High CPU usage, core saturation
• Memory Bottlenecks: Memory pressure, leak detection
• I/O Bottlenecks: Disk saturation, network congestion
• Coordination Bottlenecks: Agent communication delays
• Task Queue Bottlenecks: Queue backup, processing delays

# Analyze specific component
npx claude-flow bottleneck-analyze --component task-queue

# Full system analysis
npx claude-flow bottleneck-analyze

3. SLA Monitoring

Configure and monitor SLA metrics:

4. Anomaly Detection

Multi-model anomaly detection:

• Statistical: 3-sigma rule deviation detection
• Machine Learning: Trained anomaly detection models
• Time Series: LSTM-based temporal anomaly detection
• Behavioral: Agent behavior pattern analysis

Key Metrics

KPIs Monitored

Resource Tracking

• CPU: Current, peak, average utilization with percentiles
• Memory: Usage trends, leak detection, pressure indicators
• Network: Bandwidth utilization, latency, packet loss
• Agents: Per-agent efficiency, responsiveness, reliability

MCP Integration

// Comprehensive monitoring setup
const monitoring = {
  // Start all monitors
  async startMonitoring() {
    const [health, performance, bottlenecks] = await Promise.all([
      mcp.health_check({ components: ['swarm', 'coordination'] }),
      mcp.performance_report({ format: 'detailed', timeframe: '24h' }),
      mcp.bottleneck_analyze({})
    ]);

    return { health, performance, bottlenecks };
  },

  // Agent performance tracking
  async monitorAgents(swarmId) {
    const agents = await mcp.agent_list({ swarmId });
    const metrics = new Map();

    for (const agent of agents) {
      metrics.set(agent.id, await mcp.agent_metrics({ agentId: agent.id }));
    }

    return metrics;
  }
};

Alert Configuration

# Configure performance alerts
npx claude-flow alert-config --metric cpu_usage --threshold 80 --severity warning

# Set up anomaly detection
npx claude-flow anomaly-setup --models ["statistical", "ml", "time_series"]

# Configure notification channels
npx claude-flow notification-config --channels ["slack", "email", "webhook"]

Integration Points

Best Practices

1. Baseline Establishment: Collect baseline metrics before monitoring for anomalies
2. Alert Tuning: Start with conservative thresholds, tune based on false positive rate
3. Multi-Layer Monitoring: Monitor at system, agent, and task levels
4. Historical Analysis: Retain metrics for trend analysis and capacity planning
5. Proactive Detection: Use predictive analytics to detect issues before impact

Example: Dashboard Data Provider

// Real-time dashboard data
const dashboardData = {
  overview: {
    swarmHealth: 'healthy',
    activeAgents: 12,
    totalTasks: 1547,
    averageResponseTime: 45  // ms
  },
  performance: {
    throughput: 250,  // tasks/sec
    latency: { p50: 40, p90: 85, p99: 120 },  // ms
    errorRate: 0.02,  // percentage
    utilization: 0.72  // percentage
  },
  alerts: [],
  timestamp: Date.now()
};

Related Skills

• optimization-benchmark - Comprehensive performance benchmarking
• optimization-load-balancer - Dynamic load distribution
• optimization-resources - Resource allocation and scaling
• optimization-topology - Network topology optimization

Version History

• 1.0.0 (2026-01-02): Initial release - converted from performance-monitor agent with metrics collection, bottleneck detection, SLA monitoring, anomaly detection, and dashboard integration