Skip to content

MONITORING.md

Latest commit

 

History

History
65 lines (44 loc) · 3.4 KB

MONITORING.md

File metadata and controls

65 lines (44 loc) · 3.4 KB

Monitoring

MNM uses Prometheus for metric storage and Grafana for visualization.

Network Devices
    │
    ├── SNMP (universal) ──► SNMP Exporter ──► Prometheus ──► Grafana
    │                                               │
    └── gNMI (modern) ──────► gnmic ───────────────►│
                                                     │
                                                     ▼
                                              365-day retention
                                              10 GB size cap

Components

Prometheus — central metric store, scrapes SNMP exporter and gnmic every 30-60 seconds. Retention defaults to 365 days with a 10 GB size cap. Accessible at :8080/prometheus/. See Prometheus documentation.

Grafana — dashboard UI with two pre-built dashboards. Anonymous read-only access enabled (appliance model). Accessible at :8080/grafana/ or :3000 direct. See Grafana documentation.

SNMP Exporter — Prometheus-native SNMP poller. Polls IF-MIB (interface counters, status, errors) from configured devices. Read-only: SNMP GET/WALK only. See snmp_exporter documentation.

gnmic — gNMI streaming telemetry collector. Subscribes to OpenConfig paths (interface counters, oper-status, system state). Ready when devices have gNMI enabled. See gnmic documentation.

Dashboards

Device Dashboard (home) — per-device view with hostname dropdown. Shows interface count, up/down status, uptime, inbound/outbound traffic rates, errors, discards, and interface status table.

Network Overview — fleet-wide view showing device count, total interfaces, top talkers by traffic, and error rates across all monitored devices.

Proxmox Overview — populated by the Proxmox connector. Hypervisor node CPU/memory, top VMs by CPU/memory, full VM inventory table, ZFS pool usage with thresholds, ZFS pool health, ZFS growth over time, and physical disk SMART health. Only renders data when PROXMOX_HOST is configured.

Adding Devices to Monitoring

SNMP targets are defined in config/prometheus/prometheus.yml under the snmp_exporter job. Each target needs:

  • IP address
  • device_name label (for Grafana dropdown)

Example:

static_configs:
  - targets: ["192.0.2.5"]
    labels:
      device_name: "switch-01"

After editing, restart Prometheus: docker compose restart prometheus

Enabling gNMI on Devices

gNMI targets are configured in config/gnmic/gnmic.yml. MNM subscribes in read-only mode (subscribe, never SetRequest).

Juniper Junos — enable gNMI with:

set system services extension-service request-response grpc ssl address 0.0.0.0 port 32767

See Junos gNMI documentation.

Cisco IOS-XE — see IOS-XE gNMI configuration.

Prometheus Retention

Default: 365 days with 10 GB size cap. Configurable via PROMETHEUS_RETENTION_DAYS in .env. The size cap prevents disk exhaustion if device count grows beyond estimates. Prometheus automatically removes oldest data when either limit is reached.