Struct StaticMetricsInit 

pub struct StaticMetricsInit {
    pub stt_model: Option<String>,
    pub compaction_model: Option<String>,
    pub semantic_cache_enabled: bool,
    pub embedding_model: String,
    pub self_learning_enabled: bool,
    pub active_channel: String,
    pub token_budget: Option<u64>,
    pub compaction_threshold: Option<u32>,
    pub vault_backend: String,
    pub autosave_enabled: bool,
    pub model_name_override: Option<String>,
}

Configuration-derived fields of MetricsSnapshot that are known at agent startup and do not change during the session.

Pass this struct to AgentBuilder::with_static_metrics immediately after AgentBuilder::with_metrics to initialize all static fields in one place rather than through scattered send_modify calls in the runner.

Examples

use zeph_core::metrics::StaticMetricsInit;

let init = StaticMetricsInit {
    active_channel: "cli".to_owned(),
    ..StaticMetricsInit::default()
};

Fields

stt_model: Option<String>

STT model name (e.g. "whisper-1"). None when STT is not configured.

compaction_model: Option<String>

Model used for context compaction/summarization. None when no summary provider is set.

semantic_cache_enabled: bool

Whether the semantic response cache is enabled.

This value is also written to MetricsSnapshot::cache_enabled which is an alias for the same concept.

embedding_model: String

Embedding model name (e.g. "nomic-embed-text"). Empty when embeddings are disabled.

self_learning_enabled: bool

Whether self-learning (skill evolution) is enabled.

active_channel: String

Active I/O channel name: "cli", "telegram", "tui", "discord", "slack".

token_budget: Option<u64>

Token budget for context window. None when not configured.

compaction_threshold: Option<u32>

Token threshold that triggers soft compaction. None when not configured.

vault_backend: String

Vault backend identifier: "age", "env", or "none".

autosave_enabled: bool

Whether assistant messages are auto-saved to memory.

model_name_override: Option<String>

Override for the active model name. When Some, replaces the model name set by the builder from runtime.model_name (which may be a placeholder) with the effective model resolved from the LLM provider configuration.

Trait Implementations

impl Debug for StaticMetricsInit

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

Formats the value using the given formatter.

impl Default for StaticMetricsInit

fn default() -> StaticMetricsInit

Returns the “default value” for a type.