USER GUIDE

Using the operator surfaces

netevd has no shipped browser console for day-2 work. Operators use hook directories, the CLI, REST/Prometheus, and systemd journals. A static dashboard HTML exists in the repo for future embedding but is not the primary surface.

Mental model

Kernel / managers  →  netevd watchers  →  NetworkState
                              │
              ┌───────────────┼───────────────┐
              ▼               ▼               ▼
           hooks.d/*      REST + /metrics :9090
              │               │
              ▼               ▼
         your scripts     netevd status/list/events

Where to look

Surface	Role
/etc/netevd/*.d/	Event scripts (primary UX)
/etc/netevd/netevd.yaml	Backend, filters, API bind, metrics
netevd status\|list\|show\|events	Live inspection via API
journalctl -u netevd	Daemon + hook exit logs
:9090	Status, events, and Prometheus /metrics on <host>

Operate from CLI

1. Hooks first: add executable scripts under the event directory that matches your backend (Page guides).
2. Validate: netevd validate -c /etc/netevd/netevd.yaml
3. Observe: netevd events -f while bouncing a link (ip link set eth0 down/up).
4. Inspect state: netevd list interfaces, netevd list rules, netevd status -f json.
5. Fleet: scrape http://<host>:9090/metrics or curl http://<host>:9090/api/v1/status over an SSH tunnel.

Tips

1. Start with one routable.d logger before complex automation.
2. Use netevd events -f while bouncing a link to confirm events.
3. On multi-homed hosts, confirm policy rules with netevd list rules or ip rule.
4. Keep hooks small and executable; non-zero exits are logged and do not stop other scripts.

Next: Page-by-page guides · Workflows