Skip to main content
Real-time and historical metrics for your deployed MCP servers. Understand how your server is being used, identify bottlenecks, and spot errors before users report them.

Dashboard Overview

Access analytics in the Concierge Platform after deploying your server with concierge deploy.

Metrics Available

Throughput

Tool Popularity

See which tools are called most frequently:
Tools with high call counts but low stage transition rates may indicate the LLM is stuck:consider improving tool descriptions or adding guardrails.

Error Rates

Success Rate

Track end-to-end workflow completion:

Time-Based Usage

  • Hourly/daily/weekly request volume
  • Peak hours identification
  • Trend lines for growth tracking
  • Latency percentiles (p50, p95, p99) over time

Stage Analytics

See how users flow through your staged workflows: Identify where users drop off and optimize those stages.

Alerts

Set up alerts in the platform for:
  • Error rate exceeding a threshold
  • Response time spikes
  • Unusual traffic patterns
  • Stage drop-off rate changes