Use when configuring Prism clustering, federation, node discovery, replication, or multi-node deployments. Covers static/DNS/gossip discovery, QUIC transport, zone-aware placement, and split-brain handling.
Guide for distributed Prism deployments with horizontal scaling and high availability.
| Mode | Nodes | Consensus | Use Case |
|---|---|---|---|
| Single | 1 | None | Development, small datasets |
| Federation | 2+ | None | Read scaling, simple HA, multi-DC |
| Cluster | 3+ | Raft | Full distributed, auto-failover |
Architecture layers:
Layer 3: Cluster (Raft consensus, auto-rebalancing, leader election)
Layer 2: Federation (Query routing, result merging, health checks)
Layer 1: Single Node (Indexing, search, aggregations)
[federation.discovery]
backend = "static"
nodes = ["node1:3000", "node2:3000", "node3:3000"]
[federation.discovery]
backend = "dns"
dns_name = "prism-headless.default.svc.cluster.local"
dns_refresh_interval_secs = 30
Kubernetes headless service:
apiVersion: v1