HA / Witness

Status

GET /api/v1/ha/status — node id/role, peers, ready/health, lag.

Studio HA table: amber “Lag” when VION_HA_LAG_WARN_MS is exceeded.

Admin: POST /api/v1/ha/role (confirmed).

Optional promote

When VION_HA_FAILOVER=1, if during status aggregation all polled active peers are unhealthy, a local standby may become active in memory. A healthy active or witness does not promote. ha.role_changed is not written; process env does not change.

OPC UA server with VION_OPCUA_SERVER_MODE=standby rejects writes.

Not available yet

There is no product-level single-site HA (supervised automatic failover, cert renewal workflow, historian failover, federation).