Crate rocketmq_runtime

Expand description

Runtime substrate for the RocketMQ Rust unified thread model.

This crate standardizes how components own or borrow a Tokio runtime, how they derive service-level task scopes through RootServiceContext and ChildServiceContext, how TaskGroup and ScheduledTaskGroup track long-running and periodic tasks, how BlockingExecutor isolates short blocking work, and how ShutdownReport records verifiable shutdown evidence.

Process entrypoints own RuntimeOwner and derive a single RootServiceContext. Production libraries receive a ChildServiceContext or a narrower capability. [RuntimeContext] is a test and migration harness.

prelude documents the recommended entry path with an example that runs as a test.

Re-exports§

pub use blocking::BlockingExecutorSnapshot;
pub use blocking::BlockingKind;
pub use blocking::BlockingLane;
pub use blocking::BlockingLanePolicies;
pub use blocking::BlockingPoolPolicy;
pub use blocking::BlockingTaskSnapshot;
pub use resource_budget::BudgetCapacity;
pub use resource_budget::BudgetClass;
pub use resource_budget::BudgetDimension;
pub use resource_budget::BudgetLimit;
pub use resource_budget::BudgetRejection;
pub use resource_budget::BudgetRejectionReason;
pub use resource_budget::BudgetSnapshot;
pub use resource_budget::BudgetedItem;
pub use resource_budget::BudgetedQueue;
pub use resource_budget::DynamicBudgetKey;
pub use resource_budget::DynamicKeyAdmissionRejection;
pub use resource_budget::DynamicKeyRegistrationFailure;
pub use resource_budget::DynamicKeyRetirement;
pub use resource_budget::ForeignPermit;
pub use resource_budget::FullPolicy;
pub use resource_budget::HostMemoryFileView;
pub use resource_budget::ManagedMemoryBudget;
pub use resource_budget::ManagedMemoryPolicy;
pub use resource_budget::ManagedMemoryRequest;
pub use resource_budget::MemoryFileView;
pub use resource_budget::MemoryLimitSource;
pub use resource_budget::MonotonicClock;
pub use resource_budget::PermitRebindOutcome;
pub use resource_budget::ProcessMemoryLimit;
pub use resource_budget::QueuePushOutcome;
pub use resource_budget::QueuePushRejection;
pub use resource_budget::QueueSnapshot;
pub use resource_budget::RateLimit;
pub use resource_budget::ResourceBudgetTree;
pub use resource_budget::ResourcePermit;
pub use resource_budget::SystemMonotonicClock;
pub use service_context::ScopeId;
pub use shutdown_report::ShutdownAnnotation;
pub use shutdown_report::TaskSnapshot;
pub use signal::wait_for_signal;
pub use signal::wait_for_signal_result;
pub use task_group::TaskGroupEventCounts;
pub use task_group::TaskGroupId;
pub use task_group::TaskGroupLifecycleState;
pub use task_group::TaskId;
pub use task_group::TaskKind;
pub use task_group::TaskName;
pub use task_group::TaskResult;
pub use crate::blocking::BlockingDrainLease;
pub use crate::blocking::BlockingExecutor;
pub use crate::blocking::BlockingExecutorShutdownReport;
pub use crate::resource_budget::ResourceBudget;
pub use crate::service_context::ChildServiceContext;
pub use crate::service_context::RootServiceContext;
pub use crate::shutdown_deadline::ShutdownDeadline;
pub use crate::shutdown_report::ShutdownReport;
pub use crate::task_group::TaskGroup;

Modules§

blocking: Root-owned blocking admission with lane isolation and one global hard limit.
common: Common types and operations.
prelude: Minimal imports for owned service lifecycles, and the recommended entry path.
resource_budget: Resource budget types and operations.
service_context: Service context types and operations.
shutdown_deadline: Shutdown deadline types and operations.
shutdown_report: Shutdown report types and operations.
signal: Signal types and operations.
task: Task types and operations.
task_group: Task group types and operations.

Structs§

CriticalFailure: A durable record of one critical task failure.
CriticalFailureState: The owner-scoped authority for critical task failures.
CriticalFailureSubscription: A bounded subscription to critical failure notifications.
ExternalOperation: A static, low-cardinality label for an operation owned outside this crate.
LocalMetadataFileSystem: The production local filesystem implementation.
MetadataDeadline: An immutable absolute deadline for metadata admission and durability waits.
MetadataGeneration: A monotonically increasing generation for one metadata resource.
MetadataIoActor: A cloneable handle to the bounded metadata I/O owner.
MetadataIoActorLimitsProfile: The admission limits the metadata actor owns.
MetadataIoBlockingLaneProfile: The shared blocking lane limits in force for the metadata actor.
MetadataIoConfig: Metadata actor admission limits and legacy blocking-lane fields.
MetadataIoEffectiveProfile: A read-only view of the metadata I/O limits actually in force.
MetadataIoLegacyBlockingField: One retained actor field together with the value that applies instead.
MetadataIoObserver: Read-only access that does not retain metadata admission or target ownership.
MetadataIoPlan: Metadata actor settings that passed deterministic admission validation.
MetadataIoReceipt: A receipt for an accepted immutable snapshot.
MetadataIoResourceSnapshot: A snapshot of one logical metadata resource.
MetadataIoShutdownReport: Result of stopping admission and draining accepted work.
MetadataIoSnapshot: A point-in-time actor snapshot.
MetadataTargetIdentity: Process-local identity of one retained target history.
MetadataTargetRegistryStats: Aggregate registry occupancy without resource names or target paths.
MetadataWriteRequest: An immutable write request accepted by MetadataIoActor.
OperationContext: Cancellation, deadline, and task-class metadata for one bounded operation.
OperationWaitPolicy: Absolute deadlines for graceful completion and subsequent abort confirmation.
RuntimeBlockingKindSummaryV1: Aggregated task counts and elapsed-time information for one blocking-work kind.
RuntimeBlockingLaneSummaryV1: Aggregated state for one blocking-executor lane.
RuntimeBlockingSectionV2: Bounded blocking-executor state, including work that outlived its wait.
RuntimeConfig: Thread, driver and shutdown policy for one owned Tokio runtime.
RuntimeDiagnostics: Represents runtime diagnostics.
RuntimeDiagnosticsInputs: Component-owned inputs for the V2 view.
RuntimeDiagnosticsSnapshot: Represents runtime diagnostics snapshot.
RuntimeDiagnosticsViewOptions: Bounds applied while creating RuntimeDiagnosticsViewV1.
RuntimeDiagnosticsViewOptionsV2: Bounds applied while creating RuntimeDiagnosticsViewV2.
RuntimeDiagnosticsViewV1: Versioned, bounded, and sanitized runtime diagnostics safe for authenticated operational APIs.
RuntimeDiagnosticsViewV2: Versioned, bounded, and sanitized runtime diagnostics with explicit scopes.
RuntimeError: Stable, catalog-backed operational runtime failure.
RuntimeMetadataSectionV2: Bounded retained-metadata state.
RuntimeOwner: Owns one Tokio runtime and its root service context.
RuntimeOwnerPlan: A runtime configuration that has passed deterministic validation.
RuntimeResources: Process-wide resource capabilities owned by crate::RuntimeOwner.
RuntimeScheduleSectionV2: Bounded scheduled-work aggregates.
RuntimeShutdownSectionV2: The most recent shutdown result for one group population.
RuntimeTaskDetailV2: One bounded task detail.
RuntimeTaskDetails: A bounded detail sample that does not scan aggregate subtree statistics.
RuntimeTaskKindSummaryV1: Aggregated task counts and elapsed-time information for one task kind.
RuntimeTaskSectionV2: Bounded task counts for one group population.
ScheduledExecutionPolicy: Bounded execution policy for one scheduled task.
ScheduledTaskConfig: Name, cadence and run limit of one schedule.
ScheduledTaskGroup: Represents scheduled task group.
ScheduledTaskObserver: Reads an explicitly selected set of schedules without retaining their owner.
ScheduledTaskSnapshot: Represents scheduled task snapshot.
ServiceLifecycle: Shared process lifecycle and health-probe owner.
ServiceLifecycleConfig: Versioned runtime configuration for the process lifecycle boundary.
ServiceLifecycleTransition: One committed process state change. from = None is the initial Starting event.
ShutdownRequest: Immutable first shutdown request and its shared absolute deadline.
TaskSpawner: Narrow capability for spawning work owned by an existing TaskGroup.

Enums§

CriticalFailureKind: How a critical task stopped being healthy.
CriticalFailureRecovery: How a handled critical task failure affects the service.
DependencyReadiness: Aggregate dependency readiness, independent of maintenance and process shutdown.
DrainRequestMethods: HTTP methods that the /drainz probe route accepts.
MetadataDurability: The caller-visible completion boundary for a persistence request.
MetadataIoAdmissionOutcome: The normal admission result for one metadata write request.
MetadataIoCommitAdmissionOutcome: Result of waiting for a committed metadata snapshot.
MetadataIoCommitObservation: What a caller actually confirmed about one admitted metadata snapshot.
MetadataIoCommitOutcome: The real persistence conclusion for an accepted metadata snapshot.
MetadataIoDurabilityOutcome: The normal durable-completion result for one metadata write request.
MetadataIoOperation: A filesystem step in the atomic replacement protocol.
MetadataLimitSource: Identifies the owner of one effective metadata limit.
MetadataTargetRetirementOutcome: Result of explicitly ending a quiescent target’s history.
MissedTickPolicy: Policy for ticks that arrive while all run slots are occupied.
OperationOutcome: Why an accepted operation task finished, after its future was destroyed.
OperationWaitOutcome: Whether an operation’s task destruction was confirmed when waiting stopped.
RuntimeBlockingKindV1: Identifies a bounded category of blocking work.
RuntimeBlockingLaneV1: Identifies a bounded blocking-executor lane.
RuntimeComponent: A stable component identifier used by the sanitized diagnostics view.
RuntimeContractPolicy: A closed identifier for a deterministic runtime contract rule.
RuntimeContractViolation: A deterministic caller configuration or invariant violation.
RuntimeDiagnosticsScope: Names the population a V2 diagnostics value covers.
RuntimeErrorKind: Why a runtime operation failed, for callers that branch on the reason.
RuntimeLifecycleStateV1: Describes the lifecycle state of the root task group.
RuntimeOperation: A closed, low-cardinality runtime operation identifier.
RuntimeTaskKindV1: Identifies a bounded category of runtime-owned task.
ScheduleMode: Identifies the schedule mode state.
ScheduledTaskConcurrency: Maximum concurrent runs admitted for one bounded schedule.
ScheduledTaskControl: Identifies the scheduled task control state.
ScheduledTaskRegistrationOutcome: Describes the result of registering a scheduled task.
ServiceLifecycleState: Stable process lifecycle states used by readiness and liveness probes.
ShutdownReason: Source of the first shutdown request.

Traits§

MetadataFileSystem: Filesystem abstraction used by the metadata actor.
OperationOutcomeObserver: Observes task outcomes after future destruction.
ServiceLifecycleObserver: Observes committed state changes without owning the lifecycle.

Functions§

current_thread_committed_stack_bytes: Returns how many bytes of the calling thread’s stack are committed.

Type Aliases§

RuntimeResult: Alias for an operational runtime result.