Monitoring

nSelf includes a full monitoring stack that runs alongside your backend. It is free, included in every install, and requires no configuration.

What's included

Service	Port	Purpose
Prometheus	9090	Metrics collection
Grafana	3000 (internal)	Dashboards
Loki	3100	Log aggregation
Promtail	—	Log shipper (runs as sidecar)
Tempo	—	Distributed tracing
Alertmanager	9093	Alert routing
cAdvisor	8080	Container metrics
Node Exporter	9100	Host metrics
Postgres Exporter	9187	Database metrics
Redis Exporter	9121	Cache metrics

All monitoring services bind to 127.0.0.1. Access via Nginx proxy at monitor.<your-domain> or localhost:3000 (Grafana) during local dev.

Enable monitoring

Monitoring is enabled by default. To disable:

nself config set monitoring.enabled false
nself build  # regenerates docker-compose without monitoring stack

Accessing Grafana

nself admin open monitoring
# Opens http://localhost:3000 in your browser
# Default credentials: admin / (generated at install, see .env.secrets GRAFANA_PASSWORD)

Pre-built dashboards

Grafana includes dashboards for:

• System Overview — CPU, memory, disk, network
• Postgres — query performance, connections, cache hit ratio
• Hasura — request rate, error rate, latency
• Container resources — per-service CPU and memory
• Application logs — via Loki data source

Custom alerts

Alerts are configured in nself/alertmanager/config.yml. After editing:

nself build
nself start

Monitoring in production

The monitoring stack runs on the same VPS as your nSelf instance. For the nSelf Cloud tier, monitoring data is stored in isolated per-tenant volumes.

Troubleshooting

Prometheus not scraping a service:

nself logs prometheus --tail 50

Grafana can't connect to Loki:

nself doctor --check monitoring