Configuration

orbit.Config is bound from the modules.orbit.* subtree of your nucleus.yml (or set it directly in Go). All fields are optional.

Key (modules.orbit.*)	Type	Default	Description
prefix	string	/admin	URL path Orbit mounts under.
title	string	—	Heading shown in the UI.
environment	string	—	Label shown in the UI (e.g. production).
bootstrap_username	string	—	Admin user created on first boot.
bootstrap_email	string	—	Email for the bootstrap user.
bootstrap_password	string	—	Password for the bootstrap user; empty → a random one is generated and printed once.
auth_database	string	app default	DB alias whose handle backs admin auth + the bootstrap user (use a dedicated DB for the admin user store).
migrations_path	string	migrations	Directory the migrations view reads.
audit_max_size	int	10000	In-memory audit-log ring size.
multitenant_enabled	bool	false	Filter records by the request's resolved tenant.
multitenant_default	string	—	Default tenant when none is resolved.
multitenant_ids	[]string	—	Known tenant IDs for the selector UI.
live_exclude_patterns	[]string	—	Path patterns excluded from the live HTTP feed.
trace_url_template	string	—	External trace-explorer URL template (supports {trace_id}).
cluster_enabled	bool	false	Aggregate the live feed across nodes via a Redis relay.
cluster_redis_url	string	—	Redis URL for the live-telemetry relay.
cluster_channel	string	nucleus:admin:live:v1	Pub/sub channel for the relay.
cluster_node_id	string	runtime id	Explicit node identifier in the relay.
cluster_token	string	—	Shared secret to reject untrusted relay messages.

Example

# nucleus.yml
modules:
  orbit:
    prefix: /admin
    title: Acme Admin
    environment: production
    bootstrap_username: admin
    bootstrap_email: admin@acme.test

A note on the cluster keys

The cluster_* keys above configure the single-process Redis relay for aggregating the live feed across nodes. That is distinct from the standalone agent/server fleet model (a dedicated observability server that many agents stream to) documented under Cluster observability. Most applications need neither — the in-process panel works on its own.