Observability

A3S Gateway provides built-in Prometheus metrics, structured access logging, distributed tracing, and a dashboard API.

Dashboard API

Prop

Type

Health Response

{
  "state": "Running",
  "uptime_secs": 3600,
  "active_connections": 42,
  "total_requests": 12345
}

Prometheus Metrics

gateway_requests_total 12345
gateway_responses_total{status_class="2xx"} 11000
gateway_responses_total{status_class="3xx"} 500
gateway_responses_total{status_class="4xx"} 700
gateway_responses_total{status_class="5xx"} 145
gateway_response_bytes_total 987654321
gateway_active_connections 42
gateway_router_requests{router="api"} 8000
gateway_router_requests{router="web"} 4345
gateway_backend_requests{backend="http://127.0.0.1:8001"} 6000
gateway_backend_requests{backend="http://127.0.0.1:8002"} 6345

Grafana Queries

# Request rate
rate(gateway_requests_total[5m])

# Error rate (4xx + 5xx)
rate(gateway_responses_total{status_class=~"4xx|5xx"}[5m])

# Success rate %
rate(gateway_responses_total{status_class="2xx"}[5m])
  / rate(gateway_requests_total[5m]) * 100

# Active connections
gateway_active_connections

# Traffic by router
rate(gateway_router_requests[5m])

# Throughput (bytes/sec)
rate(gateway_response_bytes_total[5m])

Access Logging

Structured JSON access logs via the tracing crate:

{
  "timestamp": "2025-01-15T10:30:00Z",
  "client_ip": "203.0.113.42",
  "method": "POST",
  "path": "/api/v1/chat",
  "status": 200,
  "request_duration_ms": 145,
  "bytes_sent": 2048,
  "router": "api",
  "service": "api-backend",
  "backend": "http://127.0.0.1:8001"
}

Log Levels

# Via CLI
a3s-gateway --config gateway.hcl --log-level debug

# Via environment
RUST_LOG=a3s_gateway=debug a3s-gateway --config gateway.hcl

Prop

Type

Distributed Tracing

A3S Gateway propagates trace context across services:

W3C Trace Context: traceparent / tracestate headers
B3/Zipkin: X-B3-TraceId / X-B3-SpanId headers

Spans are created for request lifecycle, backend connections, and health check probes.

Observability

On this page