Struct OrchestratorCoordinator 

pub struct OrchestratorCoordinator { /* private fields */ }

Orchestration coordinator holding shared backend references.

Created by crate::app::RustvelloApp or directly via Self::new for the from_backends() FFI path. All methods are &self — the coordinator is logically immutable once built.

Implementations

impl OrchestratorCoordinator

pub async fn route_call( &self, new_invocation_id: &InvocationId, call_dto: &CallDTO, cc_args: Option<&SerializedArguments>, registration_cc: ConcurrencyControlType, index_cc: bool, runner_id: &RunnerId, ) -> RustvelloResult<RouteCallResult>

Route a call: check registration CC, create or reuse an invocation, route.

Mirrors pynenc’s BaseOrchestrator.route_call():

1. If registration_cc == Unlimited: always create a new invocation.
2. Else: query existing REGISTERED invocations with matching CC args.
   • No match → create new.
   • Match with same call_id → reuse (return existing).
   • Match with different call_id → return ReusedDifferentCall so the caller can decide (raise error or reuse).
3. For new invocations: register + persist + index CC + route.

pub async fn reroute_invocations( &self, invocation_ids: &[InvocationId], runner_id: &RunnerId, ) -> RustvelloResult<()>

Reroute a set of invocations: transition to Rerouted, then re-enqueue.

Mirrors pynenc’s BaseOrchestrator.reroute_invocations(): For each invocation: set status REROUTED → route through broker. Invalid status transitions are silently skipped (race-safe).

pub async fn trigger_loop_iteration( &self, runner_id: &RunnerId, ) -> RustvelloResult<Vec<InvocationId>>

Execute one trigger evaluation loop iteration.

Mirrors pynenc’s BaseTrigger.trigger_loop_iteration():

1. Evaluate cron conditions (time-based).
2. Process valid conditions → match triggers → fire.
3. For each fired trigger: register a new invocation + route.

Returns the list of invocation IDs created by triggers.

pub async fn check_atomic_services( &self, runner_id: &RunnerId, service_interval_minutes: f64, spread_margin_minutes: f64, runner_timeout_seconds: f64, ) -> RustvelloResult<Option<Vec<InvocationId>>>

Execute one atomic service check: coordination + trigger loop + recording.

Mirrors pynenc’s BaseRunner._check_atomic_services() flow:

1. Register heartbeat for this runner (with can_run_atomic_service=true).
2. Get active runners eligible for atomic services.
3. Check distributed coordination algorithm (time-slot allocation).
4. If authorized: run trigger_loop_iteration, record execution.
5. Return None if not authorized, Some(created_ids) if ran.

impl OrchestratorCoordinator

pub async fn get_invocations_to_run( &self, max_num_invocations: usize, runner_id: &RunnerId, config_for_task: &dyn Fn(&TaskId) -> Option<TaskConfig>, ) -> RustvelloResult<Vec<InvocationId>>

Retrieve invocations ready to run, handling blocking priority and CC.

1. Prioritise blocking invocations (those with waiters)
2. Fill remaining slots from broker queue
3. For each candidate: check CC, set PENDING or reject/reroute
4. Reroute CC-denied invocations at the end

config_for_task resolves the effective task config for CC decisions. Returns None when the task is unknown (CC is skipped).

impl OrchestratorCoordinator

pub fn new( orchestrator: Arc<dyn Orchestrator>, state_backend: Arc<dyn StateBackend>, broker: Arc<dyn Broker>, client_data_store: Arc<ClientDataStoreManager>, trigger_manager: Option<TriggerManager>, auto_purge_hours: f64, ) -> Self

Create a new coordinator from shared backend references.

pub async fn set_invocation_status( &self, invocation_id: &InvocationId, status: InvocationStatus, runner_id: &RunnerId, ) -> RustvelloResult<InvocationStatusRecord>

Atomic status transition with all side-effects, auto-resolving trigger context.

Looks up task_id and arguments from the state backend for trigger reporting. Prefer Self::set_invocation_status_with_context when the caller already has this data.

pub async fn set_invocation_status_with_context( &self, invocation_id: &InvocationId, status: InvocationStatus, runner_id: &RunnerId, task_id: &TaskId, arguments: BTreeMap<String, String>, ) -> RustvelloResult<InvocationStatusRecord>

Atomic status transition with all side-effects and explicit trigger context.

1. Atomic status transition (validates state machine)
2. If terminal: release waiters + schedule auto-purge
3. Record history in state backend
4. Notify trigger system (if configured)

pub async fn register_invocations( &self, invocations: &[(InvocationDTO, CallDTO)], runner_id: &RunnerId, ) -> RustvelloResult<()>

Register invocations with all side-effects.

1. Upsert each invocation + call in state backend
2. Register with orchestrator (sets Registered status)
3. Record history for each
4. Notify trigger system (if configured)
5. Route all through broker

pub async fn set_invocation_result( &self, invocation_id: &InvocationId, result: &str, runner_id: &RunnerId, ) -> RustvelloResult<()>

Store result and transition to Success, auto-resolving trigger context.

pub async fn set_invocation_result_with_context( &self, invocation_id: &InvocationId, result: &str, runner_id: &RunnerId, task_id: &TaskId, arguments: BTreeMap<String, String>, ) -> RustvelloResult<()>

Store a successful result and transition to Success with explicit context.

1. Store result in state backend
2. Set status to Success (includes release_waiters, auto_purge, history, trigger)
3. Notify trigger system of result (if configured)

pub async fn set_invocation_exception( &self, invocation_id: &InvocationId, error_type: &str, error_message: &str, runner_id: &RunnerId, ) -> RustvelloResult<()>

Store exception and transition to Failed, auto-resolving trigger context.

pub async fn set_invocation_exception_with_context( &self, invocation_id: &InvocationId, error_type: &str, error_message: &str, runner_id: &RunnerId, task_id: &TaskId, arguments: BTreeMap<String, String>, ) -> RustvelloResult<()>

Store an exception and transition to Failed with explicit context.

1. Store error in state backend
2. Set status to Failed (includes release_waiters, auto_purge, history, trigger)
3. Notify trigger system of failure (if configured)

pub async fn set_invocation_retry( &self, invocation_id: &InvocationId, runner_id: &RunnerId, ) -> RustvelloResult<()>

Set an invocation for retry, auto-resolving trigger context.

pub async fn set_invocation_retry_with_context( &self, invocation_id: &InvocationId, runner_id: &RunnerId, task_id: &TaskId, arguments: BTreeMap<String, String>, ) -> RustvelloResult<()>

Set an invocation for retry with explicit context.

1. Set status to Retry (via set_invocation_status_with_context)
2. Increment retry counter
3. Re-route through broker

pub async fn get_trigger_context( &self, invocation_id: &InvocationId, ) -> (TaskId, BTreeMap<String, String>)

Look up task_id and arguments for trigger context from state backend.