Enum EventKind 

#[non_exhaustive]
pub enum EventKind {
    PromptStarted {
        model: String,
        messages_in: usize,
    },
    PromptCompleted {
        model: String,
        tokens_in: Option<u64>,
        tokens_out: Option<u64>,
        response_id: Option<String>,
    },
    ToolInvoked {
        tool_name: String,
        provider_call_id: Option<String>,
        call_id: String,
        args_json: String,
        truncated: bool,
    },
    ToolCompleted {
        tool_name: String,
        provider_call_id: Option<String>,
        call_id: String,
        result: String,
        truncated: bool,
    },
    ToolSkipped {
        tool_name: String,
        call_id: String,
        reason: String,
    },
    ToolTerminated {
        tool_name: String,
        call_id: String,
        reason: String,
    },
    ContextSampled {
        message_count: usize,
        byte_size: usize,
        token_estimate: Option<u64>,
    },
    ContextCompacted {
        evicted_count: usize,
        evicted_bytes: usize,
        carry_over: bool,
        summary_bytes: usize,
    },
    MemoryDemoted {
        demoted_count: usize,
        tags: Vec<String>,
    },
    MemoryFrameWritten {
        frame_kind: String,
        frame_count_after: Option<u64>,
        bytes_written: usize,
    },
    ComposeKernelStart {
        kernel_id: String,
        skills_registered: Option<usize>,
        tools_registered: Option<usize>,
    },
    ComposeKernelShutdown {
        kernel_id: String,
        reason: String,
    },
    ComposeLoopIteration {
        kernel_id: String,
        iteration: u64,
        skill_id: Option<String>,
        confidence: Option<f64>,
    },
    ComposeSkillResolved {
        kernel_id: String,
        skill_id: String,
        applies: bool,
        delta: Option<f64>,
        confidence: Option<f64>,
    },
    ComposeRetryAttempt {
        kernel_id: String,
        target: String,
        attempt: u64,
        classification: String,
    },
    ComposeRecovery {
        kernel_id: String,
        reason: String,
        recovered: bool,
    },
}

Payload variants. Tagged on the wire as "kind": "<dotted.name>".

New variants are additive; rename or remove is a breaking change requiring a bump of SCHEMA_VERSION.

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.

PromptStarted

A prompt is about to be sent to the model provider.

Fields

model: String

Model name as declared on the agent.

messages_in: usize

Number of messages in the history at the time of the call.

PromptCompleted

A prompt finished; the model returned a completion response.

Fields

model: String

Model name as reported by the provider response (may differ from the requested model for routed providers).

tokens_in: Option<u64>

Provider-reported input tokens, if known.

tokens_out: Option<u64>

Provider-reported output tokens, if known.

response_id: Option<String>

Provider response ID, if supplied.

ToolInvoked

A tool is about to be invoked.

Fields

tool_name: String

Tool name as registered on the agent.

provider_call_id: Option<String>

Provider-supplied tool-call ID, when present.

call_id: String

Stable internal correlation ID (always present).

args_json: String

JSON-encoded arguments (possibly truncated; see truncated).

truncated: bool

true if args_json was truncated to PAYLOAD_TRUNCATE_BYTES.

ToolCompleted

A tool finished executing.

Fields

tool_name: String

Tool name (matches the paired tool.invoked).

provider_call_id: Option<String>

Provider-supplied tool-call ID, when present.

call_id: String

Stable internal correlation ID (matches the paired tool.invoked).

result: String

Tool result text (possibly truncated; see truncated).

truncated: bool

true if result was truncated to PAYLOAD_TRUNCATE_BYTES.

ToolSkipped

A previously-ToolInvoked call was skipped by a gating hook before the tool body ran. Pairs by call_id and closes the tool.invoked/tool.completed gap that would otherwise leave the invoke event orphaned.

Fields

tool_name: String

Tool name (matches the paired tool.invoked).

call_id: String

Stable internal correlation ID (matches the paired tool.invoked).

reason: String

Human-readable reason from the gate.

ToolTerminated

A previously-ToolInvoked call triggered a hook-driven termination of the agent loop. Pairs by call_id.

Fields

tool_name: String

Tool name (matches the paired tool.invoked).

call_id: String

Stable internal correlation ID (matches the paired tool.invoked).

reason: String

Human-readable reason from the hook.

ContextSampled

The active context was sampled (typically on ConversationMemory::load).

Fields

message_count: usize

Number of messages in the loaded history.

byte_size: usize

JSON byte size of the loaded history (rough size estimate).

token_estimate: Option<u64>

Optional token-count estimate. None in the default build; populated by consumers that wire a tokenizer.

ContextCompacted

A compactor fired, replacing some evicted history with a summary artifact.

Fields

evicted_count: usize

Number of messages evicted from the active context.

evicted_bytes: usize

Approximate byte size of the evicted messages.

carry_over: bool

true if the compactor produced a carry-over artifact for the next compaction cycle.

summary_bytes: usize

Byte size of the summary text written to long-term memory.

MemoryDemoted

A demotion hook moved messages to long-term storage.

Fields

demoted_count: usize

Number of messages demoted.

tags: Vec<String>

Tags applied to the demoted frames.

MemoryFrameWritten

A frame was written to the long-term store.

Fields

frame_kind: String

Frame kind as classified by the producer (e.g. "summary", "demoted").

frame_count_after: Option<u64>

Total frame count in the store after the write. None when the producer does not expose a cheap cumulative count (e.g. memvid). Consumers SHOULD NOT assume 0 means “empty store” — use this Option and treat absence as “unknown”.

bytes_written: usize

Byte size of the written frame’s text payload.

ComposeKernelStart

A rig-compose kernel became active for a conversation.

Fields

kernel_id: String

Stable kernel identifier chosen by the producer.

skills_registered: Option<usize>

Number of skills registered at startup, when known.

tools_registered: Option<usize>

Number of tools registered at startup, when known.

ComposeKernelShutdown

A rig-compose kernel stopped processing.

Fields

kernel_id: String

Stable kernel identifier chosen by the producer.

reason: String

Producer-specific shutdown reason (e.g. "normal", "error").

ComposeLoopIteration

One iteration of a rig-compose agent/kernel loop began.

Fields

kernel_id: String

Stable kernel identifier chosen by the producer.

iteration: u64

Monotonic iteration counter inside the kernel.

skill_id: Option<String>

Skill being considered or executed during this iteration.

confidence: Option<f64>

Current confidence score, when exposed by the producer.

ComposeSkillResolved

A rig-compose skill resolution completed.

Fields

kernel_id: String

Stable kernel identifier chosen by the producer.

skill_id: String

Skill identifier.

applies: bool

Whether the skill applied to the current context.

delta: Option<f64>

Confidence delta returned by the skill, when present.

confidence: Option<f64>

Post-application confidence score, when exposed by the producer. For applies = false resolutions this is the unchanged context confidence; for applies = true it reflects confidence + delta clamped to [0.0, 1.0].

ComposeRetryAttempt

A retry attempt occurred in a rig-compose dispatch or recovery path.

rig-tap does not emit this variant itself: the [crate::DispatchObserveHook] only observes the lifecycle hooks surfaced by rig-compose and rig-compose does not currently expose a per-tool retry hook. Producers with their own retry policy (custom skills, transports, or higher-level orchestrators) should emit this variant directly via crate::emit_kind so consumers receive a consistent shape.

Fields

kernel_id: String

Stable kernel identifier chosen by the producer.

target: String

Tool or operation being retried.

attempt: u64

One-based retry attempt number.

classification: String

Retry classification chosen by the producer.

ComposeRecovery

A rig-compose recovery path completed.

Fields

kernel_id: String

Stable kernel identifier chosen by the producer.

reason: String

Recovery reason or source error classification.

recovered: bool

Whether the recovery path restored normal execution.

Implementations

impl EventKind

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

Returns the wire kind discriminant for this event.

pub fn scalar_fields(&self) -> ScalarFields<'_>

Extract the per-variant scalar correlation fields that crate::emit() surfaces directly on the tracing event so that OpenTelemetry collectors and log indexers can route on them without parsing the JSON event blob.

Absent fields are returned as "" rather than Option<&str> because tracing 0.1’s static-field model does not accept Option<&str> as a Value. Consumers should filter rig_tap.<field> != "" to detect presence.

pub fn is_prompt_related(&self) -> bool

Returns true if the event is part of the prompt lifecycle (prompt.started, prompt.completed).

pub fn is_tool_related(&self) -> bool

Returns true if the event is part of the tool lifecycle (tool.invoked, tool.completed, tool.skipped, tool.terminated).

pub fn is_memory_related(&self) -> bool

Returns true if the event is related to memory and context management.

pub fn is_compose_related(&self) -> bool

Returns true if the event is related to a rig-compose kernel or agent loop.

pub fn tool_call_id(&self) -> Option<&str>

Extacts the stable call_id for tool events, if present.

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 Debug for EventKind

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

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for EventKind

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl PartialEq for EventKind

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

Tests for self and other values to be equal, and is used by ==.

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

Tests for !=. The default implementation is almost always sufficient, and should not be overridden without very good reason.

impl Serialize for EventKind

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.

impl StructuralPartialEq for EventKind