Enum EventKind 

#[non_exhaustive]
pub enum EventKind {
    SubscriberPanicked,
    RuntimeFailure,
    SubscriberOverflow,
    ShutdownRequested,
    AllStoppedWithinGrace,
    GraceExceeded,
    AttemptStarting,
    AttemptSucceeded,
    AttemptCanceled,
    AttemptFailed,
    AttemptTimedOut,
    BackoffScheduled,
    TaskAddRequested,
    TaskAdded,
    TaskAddFailed,
    TaskRemoveRequested,
    TaskRemoved,
    TaskFinished,
    ControllerRejected,
    ControllerSubmitted,
    ControllerSlotTransition,
}

Describes what happened in the runtime.

Every event has seq, at, and kind. Variant docs list only the additional fields normally set by the runtime. Include a wildcard arm when matching because new event kinds may be added.

Variants (Non-exhaustive)

Non-exhaustive enums could have additional variants added in future. Therefore, when matching against variants of non-exhaustive enums, an extra wildcard arm must be added to account for any future variants.

SubscriberPanicked

A subscriber panicked while processing an event.

Sets:

• task: subscriber name
• reason: panic info/message
• at: wall-clock timestamp
• seq: process-local sequence

RuntimeFailure

An internal runtime component failed.

This includes a caught panic or a worker that did not join cleanly.

Sets:

• task: runtime component name
• reason: diagnostic failure details
• at: wall-clock timestamp
• seq: process-local sequence

SubscriberOverflow

An event was lost because a subscriber path fell behind or closed.

Sets:

• task: subscriber name (or the internal consumer that lagged)
• reason: "full", "closed", or "lagged(n)"
• at: wall-clock timestamp
• seq: process-local sequence

ShutdownRequested

Shutdown was requested.

This can come from an OS signal or an explicit runtime shutdown request.

Sets:

• at: wall-clock timestamp
• seq: process-local sequence

AllStoppedWithinGrace

All tasks stopped within configured grace period.

Sets:

• at: wall-clock timestamp
• seq: process-local sequence

GraceExceeded

Grace period exceeded; some tasks did not stop in time.

Sets:

• at: wall-clock timestamp
• seq: process-local sequence

AttemptStarting

A task attempt is starting.

Sets:

• id: task run identity
• task: task name
• attempt: attempt number (1-based for this task run)
• at: wall-clock timestamp
• seq: process-local sequence

AttemptSucceeded

A task attempt returned Ok(()).

This is an attempt result, not always the final task result. Under RestartPolicy::Always, another attempt follows.

Sets:

• id: task run identity
• task: task name
• attempt: attempt number
• duration_ms: attempt duration

AttemptCanceled

Task attempt returned TaskError::Canceled.

Sets:

• id: task run identity
• task: task name
• attempt: attempt number
• duration_ms: attempt duration

AttemptFailed

A task attempt returned a failure.

This includes retryable failures, fatal errors, task-returned timeouts, and panics caught while running user code. A configured per-attempt deadline instead emits AttemptTimedOut. A later event shows whether Taskvisor retries or reaches a terminal state.

Sets:

• id: task run identity
• task: task name
• attempt: attempt number
• duration_ms: attempt duration
• reason: error message
• exit_code: process-like exit code, when available

AttemptTimedOut

Task exceeded its configured timeout for this attempt.

Sets:

• id: task run identity
• task: task name
• attempt: attempt number
• timeout_ms: configured timeout
• duration_ms: elapsed attempt duration

BackoffScheduled

The next attempt was scheduled after success or failure.

Sets:

• id: task run identity
• task: task name
• attempt: previous attempt number
• delay_ms: delay before the next attempt (ms)
• backoff_source: Success or Failure
• reason: last failure message (only for failure-driven backoff)
• at: wall-clock timestamp
• seq: process-local sequence

TaskAddRequested

An add request was published before Taskvisor processed it.

This does not confirm admission. For an all-or-nothing batch, Taskvisor publishes one request event per item before it sends the whole batch command.

Sets:

• id: task run identity (pre-allocated for this add request)
• task: logical task name
• at: wall-clock timestamp
• seq: process-local sequence

TaskAdded

A task was registered and its managed runner was spawned.

Sets:

• id: task run identity
• task: task name
• at: wall-clock timestamp
• seq: process-local sequence

TaskAddFailed

A task was not added because its name conflicted or its all-or-nothing batch was rejected.

No task runner is spawned for a rejected dynamic add. If an all-or-nothing batch is rejected, no task runner is spawned for any item.

Sets:

• id: task run identity of the rejected add request
• task: task name
• outcome_kind: TaskOutcomeKind::Rejected
• rejection_kind: RejectionKind::AlreadyExists or RejectionKind::BatchRejected
• reason: diagnostic rejection details
• at: wall-clock timestamp
• seq: process-local sequence

TaskRemoveRequested

A remove request was published before Taskvisor completed it.

This is not proof of removal. Use the management method’s result or a waiter when you need a reliable answer.

Sets:

• id: task run identity
• task: task name, when known
• reason: optional removal reason
• at: wall-clock timestamp
• seq: process-local sequence

TaskRemoved

Task was removed from the supervisor after terminal cleanup.

Sets:

• id: task run identity
• task: task name
• at: wall-clock timestamp
• seq: process-local sequence

TaskFinished

A registered task reached its final outcome and will not start another attempt.

This is emitted once after the task runner is joined and before TaskRemoved. It also covers force-abort and an internal runner panic, even when no attempt-level terminal event exists.

Sets:

• id: task run identity
• task: task name
• outcome_kind: stable machine-readable final category
• reason: optional diagnostic detail; never parse it as schema
• exit_code: process-like exit code, when available
• at: wall-clock timestamp
• seq: process-local sequence

ControllerRejected

Available on crate feature controller only.

The controller rejected a submission.

Sets:

• task: slot name, when known
• id: the rejected submission’s TaskId
• outcome_kind: TaskOutcomeKind::Rejected
• rejection_kind: stable machine-readable rejection category
• reason: readable rejection details

ControllerSubmitted

Available on crate feature controller only.

The controller accepted a submission.

The task may still be queued or waiting for runtime registration. This event does not mean that the task body has started.

Sets:

• task: slot name
• id: the submission’s TaskId
• reason: a readable admission summary, e.g. admission=Queue status=admitting or started_from_queue depth=N (exact text is diagnostic, not a stable contract)

ControllerSlotTransition

Available on crate feature controller only.

A controller slot changed state.

Sets:

• task: slot name
• reason: readable transition text; it is not a stable machine contract

Implementations

impl EventKind

pub fn as_label(&self) -> &'static str

Returns a stable machine-readable label for logs and metrics.

The label is the snake_case form of the variant name. Use it as an event name in tracing or as a metrics label value.

EventKind::AttemptStarting
          │ as_label()
          ▼
    "attempt_starting"
       ├── log field:    event="attempt_starting"
       └── metric label: event="attempt_starting"

use taskvisor::EventKind;

assert_eq!(EventKind::AttemptStarting.as_label(), "attempt_starting");
assert_eq!(EventKind::BackoffScheduled.as_label(), "backoff_scheduled");

Trait Implementations

impl Clone for EventKind

fn clone(&self) -> EventKind

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Copy for EventKind

impl Debug for EventKind

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl Eq for EventKind

impl PartialEq for EventKind

fn eq(&self, other: &EventKind) -> bool

Equality operator ==.

fn ne(&self, other: &Rhs) -> bool

Inequality operator !=.

impl StructuralPartialEq for EventKind