MCP App: Server Metrics Dashboard — Tool usage and performance monitor

[jongalloway]

Summary

Create a Server Metrics Dashboard MCP App for monitoring tool usage, performance, and cache effectiveness in real-time.

MCP Apps Pattern: Real-time monitoring

Understanding how the MCP server is performing — which tools are called most, response times, cache hit rates, error rates — requires continuous visibility. A dashboard with charts and gauges updates as the server is used, without requiring "what's the status now?" prompts.

Design

• Tool usage chart: Bar chart showing invocation counts per tool
• Performance gauges: Average response time per tool category (SDK, project, package, etc.)
• Cache metrics: Hit/miss ratio gauges for SDK info, templates, and other cached resources
• Error rate: Success vs. failure counts with error categorization
• Auto-refresh: Periodically poll DotnetServerMetrics to update display

Data Sources

• DotnetServerMetrics tool with Get action — per-tool invocation counts, durations, success/failure rates
• DotnetSdk tool with CacheMetrics action — cache hit rates, TTL, and performance data
• CacheMetrics class — hits, misses, average latency

Resource URI

ui://dotnet-mcp/server-metrics

Priority

Medium — good real-time monitoring showcase, low effort since data is already structured JSON.