How-to guides

Task-focused recipes. Each page is "how do I do X?", not "what is X?". For the conceptual side, read the Concepts section.

Operating a cluster

Run in production — sizing, network surface, secrets, configuration reference, logging, shutdown.
Configure authentication — TOFU versus strict, provisioning the trust store, key rotation.
Manage node keys — create a node's keys, read its fingerprint, and share public keys so peers trust each other.
Observe a cluster — the metrics catalogue, wiring a Prometheus or OTLP exporter, alerting.
Troubleshoot — symptom-cause-fix tables for boot failures, authentication, cluster membership, GC.

Migrate connections — RFC 9000 §9 path rebind via migrate_peer/1,2, including a small watchdog recipe.
Route through a relay — wiring an external relay or tunnel adapter for peers that cannot reach each other directly.
Run the tests — local CT and EUnit suites, the gated soak suite, the docker auth integration suite.
Partition state across nodes — route keys to owners with place/1 and hand off on ownership events.
Schedule durable jobs — run work at a future time with remind/3, surviving the scheduling node's death.
Share replicated state — a cluster-wide config/flags map with barrel_p2p_map: put, get, subscribe, validate, tune GC.