Each process exposes /metrics:

lattice_up
lattice_dhcp_allocations_total / lattice_dhcp_naks_total
lattice_dhcp_offer_latency_seconds
lattice_db_query_total
lattice_pool_free, lattice_outbox_lag
lattice_dns_updates_total{result}

Logs are JSON slog: node, xid, mac, ip, subnet_id, msg_type.

API and worker /metrics are open on loopback. Scrapes from another host need Authorization: Bearer lattice_live_… (readonly is enough). dhcpd /metrics is loopback-only. Fleet exposes GET /metrics on the Fleet process (loopback open; Bearer off-loopback).

Alert on /readyz != 200, outbox lag climbing, pool free ~ 0, node heartbeat stale.

Ship logs and webhooks: Observability.