Monitoring & Metrics

Prometheus Integration

Burgonet Gateway provides built-in Prometheus metrics endpoint for monitoring and alerting. The metrics are exposed on the configured host and port (default: 127.0.0.1:6192).

Configuration

The Prometheus endpoint can be configured in conf.yml:

prometheus_host: 127.0.0.1  # Host to expose metrics on
prometheus_port: 6192       # Port for metrics endpoint

Available Metrics

The gateway exposes the following metrics:

• burgonet_requests_total (counter): Total number of requests processed
• burgonet_input_tokens_total (counter): Total number of input tokens processed
• burgonet_output_tokens_total (counter): Total number of output tokens generated

Example Prometheus Queries

1. Requests per minute:

rate(burgonet_requests_total[1m])

1. Input tokens per hour:

rate(burgonet_input_tokens_total[1h])

1. Output tokens per day:

rate(burgonet_output_tokens_total[1d])

1. Token ratio (output/input):

burgonet_output_tokens_total / burgonet_input_tokens_total

Grafana Dashboard

A sample Grafana dashboard is available in the docs/grafana directory. It includes:

• Request rate over time
• Token usage trends
• Error rates
• System resource utilization

Alerting Examples

1. High request rate alert:

- alert: HighRequestRate
  expr: rate(burgonet_requests_total[5m]) > 1000
  for: 10m
  labels:
    severity: critical
  annotations:
    summary: "High request rate detected"
    description: "Request rate is above 1000 req/min for 10 minutes"

1. Token quota warning:

- alert: TokenQuotaWarning  
  expr: burgonet_input_tokens_total % 1000000 > 900000
  for: 5m
  labels:
    severity: warning
  annotations:
    summary: "Token quota nearing limit"
    description: "Token usage is over 90% of quota"

Best Practices

• Monitor request rates and token usage trends
• Set alerts for quota thresholds
• Track token ratios to detect anomalies
• Correlate metrics with system resource usage