Production Deployment

Minimum production topology

Role	Node count	Notes
INFRA	1–3	Prometheus / Grafana / DNS / package repo
ETCD	3 or 5	DCS (must be odd)
PGSQL	3+	1 primary + 2 replicas (satisfies HA)

Steps

1. Plan — read architecture planning to decide node count, network zones, VIP strategy
2. Prepare resources — follow the resource checklist for hosts, SSH, disks, firewalls
3. Deploy — production install
4. Validate in a sandbox — use the sandbox on Vagrant / VMs first
5. Provision infrastructure (optional) — Vagrant or Terraform

Recommended templates

Reuse one of the production-ready templates:

• trio.yml — three-node HA
• full.yml — full production template
• safe.yml — cross-AZ safety template
• See config templates

Pre-launch checklist

• [ ] etcd — all three nodes healthy
• [ ] PG replication lag < 1s
• [ ] pgBackRest — first full backup complete
• [ ] Every Grafana dashboard has data
• [ ] Alert webhook verified end-to-end
• [ ] Performed one patronictl switchover drill
• [ ] Firewall exposes only the business ports
• [ ] All default passwords rotated

Going deeper

• Full deployment manual: /docs/deploy/