Status
Component health
Probes the marketing edge and (when the appliance tunnel is up) the four control-plane surfaces. See the Production runbook for rollback commands.
Operator runbook
Each control-plane component has a liveness probe and a single-line rollback. Open the runbook for the full commands, including pre-flight checks and on-call escalation.
- API / BFF — GET /v1/ready → 200 { ready: true } when Kumo is reachable. Rollback:
systemctl restart postmta-hosted-api.service· Runbook - SMTP proxy — STARTTLS on :587 with PLAIN auth via the panel-issued credentials. /v1/ready echoes SMTP readiness. Rollback:
systemctl restart postmta-hosted-smtp.service· Runbook - Kumo + Incus — kumod liveness at :8000/api/check-liveness/v1; per-node health via GET /v1/nodes/:id/health. Rollback:
systemctl restart postmta-hosted-kumo.service && incus restart· Runbook - Supabase (Postgres, Auth, Realtime, Storage) — Realtime channel on wallets; auth via Supabase JWT; migrations apply on `make appliance-up`. Rollback:
cd supabase && supabase db reset && make appliance-up· Runbook - Caddy edge — Reverse-proxies SSR :4321 + API :4000; never proxies Kumo admin directly. Headers set per-route. Rollback:
systemctl reload caddy· Runbook
If status is down, follow the Production runbook or open the operator panel for tenant-level health probes.
Incident history
We're a fresh deployment — no historical incidents yet.