Struct MultiWaypoint 

pub struct MultiWaypoint {
    pub databases: Vec<NamedDatabaseConfig>,
}

Multi-database orchestration entry point.

Fields

databases: Vec<NamedDatabaseConfig>

List of all database configurations to orchestrate.

Implementations

impl MultiWaypoint

pub fn execution_order(databases: &[NamedDatabaseConfig]) -> Result<Vec<String>>

Determine execution order based on depends_on relationships (Kahn’s algorithm).

Databases that are ready at the same moment run in declaration order — the order they appear in [[databases]] — which makes the result deterministic.

The ready set used to be seeded by iterating a HashMap, whose order is randomly seeded per process, so three independent databases migrated in a different order on almost every run. Any topological order is correct, but a varying one makes --fail-fast leave a different subset migrated each time, and makes staging and production disagree for no reason.

Uses borrowed &str references internally to avoid cloning database names during the topological sort; only clones into owned Strings for the output.

pub async fn connect( databases: &[NamedDatabaseConfig], filter: Option<&str>, ) -> Result<HashMap<String, DbClient>>

Connect to all databases (or a filtered subset). The engine for each database is auto-detected from the URL scheme — mixed PG/MySQL configs are fully supported here.

Uses built-in defaults for the global config sections. Prefer Self::connect_inheriting so that top-level [database] transport settings apply.

pub async fn connect_inheriting( databases: &[NamedDatabaseConfig], filter: Option<&str>, parent: &WaypointConfig, ) -> Result<HashMap<String, DbClient>>

Like Self::connect, but inheriting global sections from parent.

pub async fn migrate( databases: &[NamedDatabaseConfig], clients: &HashMap<String, DbClient>, order: &[String], target_version: Option<&str>, fail_fast: bool, ) -> Result<MultiResult>

Run migrate on all databases in dependency order.

pub async fn migrate_with_options( databases: &[NamedDatabaseConfig], clients: &HashMap<String, DbClient>, order: &[String], target_version: Option<&str>, fail_fast: bool, force: bool, ) -> Result<MultiResult>

Run migrate on all databases in dependency order with the force flag for overriding DANGER safety verdicts on PostgreSQL.

Uses built-in defaults for the global config sections. Prefer Self::migrate_inheriting so that top-level [safety], [preflight], [guards] and [reversals] policy applies.

pub async fn migrate_inheriting( databases: &[NamedDatabaseConfig], clients: &HashMap<String, DbClient>, order: &[String], target_version: Option<&str>, fail_fast: bool, force: bool, parent: &WaypointConfig, ) -> Result<MultiResult>

Like Self::migrate_with_options, but inheriting global config sections from parent.

pub async fn info( databases: &[NamedDatabaseConfig], clients: &HashMap<String, DbClient>, order: &[String], ) -> Result<HashMap<String, Vec<MigrationInfo>>>

Run info on all databases in dependency order.

pub async fn info_inheriting( databases: &[NamedDatabaseConfig], clients: &HashMap<String, DbClient>, order: &[String], parent: &WaypointConfig, ) -> Result<HashMap<String, Vec<MigrationInfo>>>

Like Self::info, but inheriting global config sections from parent.