Configure cluster peers

Preview only — this is a demo install (SC_READONLY=1).
The form below shows the configuration surface a real operator would use. Submissions are blocked. On your own install, the form persists to ~/.safecadence/cluster.env.

SC_HA_MODE

Which HA architecture this node participates in.

Node name

Unique identifier for this host in cluster status output.

Architecture A — shared stores

Redis URL

Coordinates the active-node lease. redis:// or rediss://.

Cluster peers

Comma-separated host[:port] list of OTHER nodes for healthz aggregation.

Database URL

Postgres connection string. Required for true HA (SQLite can't replicate).

S3 bucket

Shared bucket for reports/evidence/attachments.

S3 endpoint

Leave blank for AWS S3; set for MinIO / Wasabi / Backblaze.

Architecture B — peer-to-peer sync

Peer host

Hostname / IP of the OTHER node. Required when SC_HA_MODE=peer-sync.

Peer port

Port the OTHER node listens on for incoming events.

Local listen host

Interface this node binds for incoming peer events. Default 0.0.0.0.

Local listen port

Port this node binds for incoming peer events. Default 8767.

Shared HMAC secret

≥ 24 chars; same value on both nodes. Generate via `openssl rand -hex 32`.

How this works

Settings persist to ~/.safecadence/cluster.env (mode 0600, owner-only). No root required.
The systemd unit sources that file via EnvironmentFile=-~/.safecadence/cluster.env (the leading - means 'ignore if missing').
Changes take effect on the next service restart: systemctl restart safecadence.
Secrets are stored encrypted at rest on disk; the form displays only the last 4 characters and accepts blank as 'keep current'.
Full HA deployment guide: docs/HA_DEPLOYMENT.md