StockSpark MCP Server

const { performanceData } = require('../utils/performance'); const performanceTools = [ { name: "get_mcp_performance", description: `Check MCP performance metrics with objective benchmarks and optimization opportunities When to use: Monitor system health, identify bottlenecks, optimize workflows. This tool is for developers & system admins only. Only use when specifically asked for. Returns: Performance stats with benchmarks, error rates, and specific recommendations Benchmark context: API calls <1s (excellent), 1-2s (good), 2-5s (acceptable), >5s (needs optimization)`, inputSchema: { type: "object", properties: { detailed: { type: "boolean", description: "Include detailed operation breakdown and timing analysis", default: false } } }, handler: async (params = {}) => { const recent = performanceData.operations.slice(-50); const last20 = recent.slice(-20); if (recent.length === 0) { return { status: "no_data", message: "No operations tracked yet - performance monitoring starts after first tool use", recommendations: ["Start using MCP tools to collect performance data"] }; } const byTool = {}; // Calculate comprehensive stats recent.forEach(op => { if (!byTool[op.tool]) { byTool[op.tool] = { total: 0, count: 0, errors: 0, operations: [], minTime: Infinity, maxTime: 0 }; } byTool[op.tool].total += op.duration; byTool[op.tool].count++; byTool[op.tool].operations.push(op.duration); byTool[op.tool].minTime = Math.min(byTool[op.tool].minTime, op.duration); byTool[op.tool].maxTime = Math.max(byTool[op.tool].maxTime, op.duration); if (!op.success) byTool[op.tool].errors++; }); // Performance classification thresholds (based on API best practices) const classifyPerformance = (avgMs) => { if (avgMs < 1000) return { level: "excellent", color: "🟢" }; if (avgMs < 2000) return { level: "good", color: "🟡" }; if (avgMs < 5000) return { level: "acceptable", color: "🟠" }; return { level: "needs_optimization", color: "🔴" }; }; const stats = Object.entries(byTool).map(([tool, data]) => { const avgMs = Math.round(data.total / data.count); const performance = classifyPerformance(avgMs); return { tool, avg_ms: avgMs, min_ms: data.minTime, max_ms: data.maxTime, calls: data.count, errors: data.errors, error_rate: ((data.errors / data.count) * 100).toFixed(1) + '%', performance_level: performance.level, indicator: performance.color }; }).sort((a, b) => b.avg_ms - a.avg_ms); // Generate objective recommendations const recommendations = []; const optimizations = []; // Analyze performance patterns const slowTools = stats.filter(s => s.avg_ms > 5000); const errorProneTools = stats.filter(s => s.errors > 0); const frequentTools = stats.filter(s => s.calls >= 10); // Specific optimization recommendations slowTools.forEach(tool => { if (tool.tool.includes('upload_vehicle_images')) { optimizations.push({ tool: tool.tool, issue: `${tool.avg_ms}ms average (${tool.performance_level})`, solution: "Use upload_vehicle_images_claude for Claude UI images (70% faster)", impact: "high" }); } else if (tool.tool.includes('get_vehicle') || tool.tool.includes('list_vehicles')) { optimizations.push({ tool: tool.tool, issue: `${tool.avg_ms}ms average (${tool.performance_level})`, solution: "Consider adding filters to reduce dataset size or implement caching", impact: "medium" }); } else if (tool.tool.includes('analytics') || tool.tool.includes('underperforming')) { optimizations.push({ tool: tool.tool, issue: `${tool.avg_ms}ms average (${tool.performance_level})`, solution: "Analytics tools process large datasets - use smaller limit parameters", impact: "medium" }); } else { optimizations.push({ tool: tool.tool, issue: `${tool.avg_ms}ms average (${tool.performance_level})`, solution: "Check network connectivity and API endpoint health", impact: "unknown" }); } }); errorProneTools.forEach(tool => { optimizations.push({ tool: tool.tool, issue: `${tool.error_rate} error rate (${tool.errors}/${tool.calls} calls)`, solution: "Review error logs for authentication, validation, or data issues", impact: "high" }); }); frequentTools.forEach(tool => { if (tool.tool.includes('get_vehicle') && tool.calls >= 20) { optimizations.push({ tool: tool.tool, issue: `High frequency: ${tool.calls} calls`, solution: "Consider batching requests or caching vehicle data", impact: "medium" }); } }); // Overall system health assessment const totalErrors = recent.filter(op => !op.success).length; const errorRate = (totalErrors / recent.length * 100).toFixed(1); const avgResponseTime = Math.round(recent.reduce((sum, op) => sum + op.duration, 0) / recent.length); const systemHealth = classifyPerformance(avgResponseTime); // Business impact insights const businessInsights = []; if (parseFloat(errorRate) > 5) { businessInsights.push("High error rate may impact user experience and workflow completion"); } if (avgResponseTime > 3000) { businessInsights.push("Slow response times may delay vehicle processing and reduce productivity"); } if (optimizations.filter(o => o.impact === "high").length > 0) { businessInsights.push("High-impact optimizations available - implementing them could significantly improve performance"); } const result = { system_health: { status: systemHealth.level, indicator: systemHealth.color, avg_response_time_ms: avgResponseTime, error_rate: errorRate + '%', total_operations: recent.length, time_window: "last 50 operations" }, tool_performance: stats.slice(0, 10), // Top 10 by response time optimizations: optimizations.slice(0, 5), // Top 5 recommendations business_impact: businessInsights }; // Add detailed breakdown if requested if (params.detailed) { result.detailed_analysis = { performance_distribution: { excellent: stats.filter(s => s.performance_level === "excellent").length, good: stats.filter(s => s.performance_level === "good").length, acceptable: stats.filter(s => s.performance_level === "acceptable").length, needs_optimization: stats.filter(s => s.performance_level === "needs_optimization").length }, recent_trend: { operations: last20.length, avg_time: Math.round(last20.reduce((sum, op) => sum + op.duration, 0) / last20.length), errors: last20.filter(op => !op.success).length }, top_slow_operations: recent .sort((a, b) => b.duration - a.duration) .slice(0, 5) .map(op => ({ tool: op.tool, duration_ms: op.duration, success: op.success, timestamp: new Date(op.timestamp).toISOString() })) }; } // Return in MCP format for proper agent communication return { content: [ { type: 'text', text: JSON.stringify(result, null, 2) } ] }; } } ]; module.exports = { performanceTools };