Skip to main content

Monitoring and Logs

Comprehensive monitoring and logging are essential for maintaining secure and reliable MCP server deployments. Helmet provides real-time visibility into all server operations.

Accessing Server Logs​

From the Servers Dashboard​

  1. Navigate to the Servers section
  2. Find your deployed server
  3. Click on the server row or menu actions
  4. Select "View Logs"

Log Stream Interface​

The log viewer provides:

  • Real-time log streaming
  • Historical log browsing
  • Search and filtering
  • Export capabilities

Understanding Log Entries​

Log Structure​

Each log entry contains:

  • Timestamp: When the event occurred
  • Event Type: Category of log entry
  • Direction: Request flow (client->server or server->client)
  • Method: MCP method being called
  • Session ID: Unique identifier for the session
  • Metadata: Additional context and details

Event Types​

Session Events​

  • session_start: New client connection established
  • session_end: Client disconnection
  • proxy_close: Proxy shutdown

Request/Response Logs​

  • request: Incoming client requests
  • response: Server responses
  • error: Failed operations

System Events​

  • initialization: Server startup
  • configuration: Settings changes
  • performance: Metrics and statistics

Log Entry Details​

Request Logs​

Request entries show:

{
  "event": "request",
  "method": "tools/call",
  "params": {
    "name": "query_database",
    "arguments": {
      "query": "SELECT * FROM users"
    }
  },
  "timestamp": "2024-06-04T10:15:32Z"
}

Response Logs​

Response entries include:

{
  "event": "response",
  "result": {
    "rows": 42,
    "execution_time": "125ms"
  },
  "error": null,
  "timestamp": "2024-06-04T10:15:33Z"
}

Error Logs​

Error entries contain:

{
  "event": "error",
  "error": {
    "code": -32001,
    "message": "Tool access denied by policy",
    "data": {
      "type": "policy_block",
      "policy_id": "sec-001"
    }
  }
}

Monitoring Features​

Coming Soon!

Troubleshooting​

Use logs to diagnose:

  • Connection failures
  • Authentication issues
  • Tool execution errors
  • Performance problems
  • Policy violations

Security Monitoring​

Watch for security events:

  • Failed authentication attempts
  • Policy blocks
  • Unusual access patterns
  • Suspicious tool usage
  • Data access anomalies

Troubleshooting with Logs​

Common Scenarios​

High Error Rates

  1. Filter by error events
  2. Identify error patterns
  3. Check error messages
  4. Review recent changes
  5. Implement fixes

Performance Issues

  1. Analyze response times
  2. Check resource metrics
  3. Identify slow tools
  4. Review query patterns
  5. Optimize bottlenecks

Security Incidents

  1. Search for policy blocks
  2. Review access patterns
  3. Check authentication logs
  4. Identify anomalies
  5. Take corrective action