metrics

Metrics Demo Example

This example demonstrates how to use the metrics system in hypermcp to track server performance and usage statistics.

Features Demonstrated

• Tool invocation tracking
• Cache hit/miss tracking
• Metrics retrieval via tool
• Periodic metrics logging
• Final metrics on shutdown
• Cache-specific metrics

Running the Example

go run main.go

Testing with MCP Inspector

Use the MCP Inspector to test the tools:

1. Call process_data with different keys to populate the cache:

   {"key": "test1"}
   {"key": "test2"}
   {"key": "test1"}  // This will hit the cache
   

2. Call get_metrics to see current statistics:

   {}
   

What to Observe

• First call to process_data with a key will be a cache miss
• Subsequent calls with the same key (within 2 minutes) will be cache hits
• get_metrics shows real-time statistics
• Every 30 seconds, metrics are logged to console
• On shutdown (Ctrl+C), final metrics and cache statistics are displayed

Metrics Tracked

• Uptime: How long the server has been running
• Tool Invocations: Number of times tools have been called
• Resource Reads: Number of resource read operations
• Cache Hits: Number of successful cache retrievals
• Cache Misses: Number of cache misses
• Cache Hit Rate: Percentage of cache hits vs total cache operations
• Errors: Number of errors encountered

Cache Metrics

The example also demonstrates accessing Ristretto cache metrics:

• Hits and misses
• Hit ratio
• Other internal cache statistics