Monitoreo
Métricas, alertas y observabilidad
DuneStack expone métricas en formato Prometheus y tiene observabilidad integrada vía Eye of Ibad. No necesitas instalar Grafana ni Prometheus por separado.
Métricas integradas
Cada componente expone métricas que Fremen recolecta y almacena en Sietch. Eye of Ibad las visualiza en tiempo real.
bash
# Ver métricas del cluster
dune metrics
# COMPONENT CPU RAM REQ/S ERRORS
# Spice 2% 45MB — 0
# Spacing Guild 8% 64MB 12,430 0.01%
# Fremen (avg) 3% 18MB — 0
# Sandworm 1% 32MB — 0
# Sietch 4% 90MB 847/s 0
# Exportar en formato Prometheus
curl http://localhost:7430/metrics
# dune_spice_raft_term 42
# dune_gateway_requests_total 1247891
# dune_fremen_cpu_percent{node="prod-01"} 12.4Alertas
toml
# /etc/dune/alerts.toml
[[alert]]
name = "high_cpu"
condition = "fremen_cpu_percent > 80"
for = "5m"
action = "webhook"
webhook_url = "https://hooks.slack.com/..."
[[alert]]
name = "node_down"
condition = "fremen_heartbeat_age > 30s"
action = "webhook"
webhook_url = "https://hooks.slack.com/..."
[[alert]]
name = "disk_full"
condition = "fremen_disk_percent > 90"
action = "webhook"
webhook_url = "https://hooks.slack.com/..."