Monitoring

Graphmind exposes Prometheus metrics, a health check endpoint, and structured logging for observability.

Health Check

curl http://localhost:8080/api/status

{
  "status": "healthy",
  "version": "0.6.1",
  "storage": {
    "nodes": 1042,
    "edges": 3891
  }
}

Use this endpoint for load balancer health checks, Docker HEALTHCHECK, or Kubernetes liveness probes.

RESP-based health check:

redis-cli -p 6379 PING
# PONG

Prometheus Metrics

Graphmind exposes metrics at GET /metrics in Prometheus text format.

curl http://localhost:8080/metrics

Available Metrics

Metric	Type	Description
graphmind_nodes_total	Gauge	Total number of nodes
graphmind_edges_total	Gauge	Total number of edges
graphmind_queries_total	Counter	Total queries executed (labels: type=read|write)
graphmind_query_duration_ms	Histogram	Query execution time in ms (labels: type=read|write)
graphmind_resp_connections_active	Gauge	Currently active RESP connections
graphmind_resp_connections_total	Counter	Total RESP connections since startup
graphmind_script_statements_total	Counter	Total script statements executed
graphmind_script_errors_total	Counter	Script statement errors

Prometheus Configuration

Add Graphmind as a scrape target in prometheus.yml:

scrape_configs:
  - job_name: 'graphmind'
    scrape_interval: 15s
    static_configs:
      - targets: ['localhost:8080']

Grafana Dashboard

With Prometheus scraping metrics, you can create Grafana dashboards. Useful panels:

• Query rate: rate(graphmind_queries_total[5m])
• Query latency (p95): histogram_quantile(0.95, rate(graphmind_query_duration_ms_bucket[5m]))
• Active connections: graphmind_resp_connections_active
• Graph size: graphmind_nodes_total and graphmind_edges_total
• Write vs read ratio: rate(graphmind_queries_total{type="write"}[5m]) / rate(graphmind_queries_total[5m])

Logging

Graphmind uses structured logging via the tracing crate. Control the log level with:

# Environment variable
RUST_LOG=info graphmind

# Config file
[logging]
level = "debug"

# CLI flag
graphmind --log-level debug

Log levels: error, warn, info, debug, trace

Log output

Logs are written to stdout in a human-readable format by default. For JSON-structured output suitable for log aggregation (ELK, Loki, etc.), set:

RUST_LOG=info graphmind

What gets logged

Level	Events
error	Failed queries, persistence errors, startup failures
warn	Authentication failures, deprecated features, config issues
info	Server start/stop, recovery, connection events
debug	Individual query execution, plan details
trace	RESP protocol frames, operator-level execution

Docker Health Check

The Graphmind Docker image includes a built-in health check:

# Check container health
docker inspect --format='{{.State.Health.Status}}' graphmind

The health check sends a RESP PING every 30 seconds and expects PONG.

Kubernetes Probes

livenessProbe:
  httpGet:
    path: /api/status
    port: 8080
  initialDelaySeconds: 10
  periodSeconds: 30

readinessProbe:
  httpGet:
    path: /api/status
    port: 8080
  initialDelaySeconds: 5
  periodSeconds: 10