Crate llm_analytics_hub 

LLM Analytics Hub

Centralized analytics hub for the LLM ecosystem, providing comprehensive data models and schemas for telemetry, security, cost, and governance monitoring.

Overview

This crate provides:

• Event Schemas: Unified event schema for telemetry, security, cost, and governance
• Metrics Models: Time-window aggregations with statistical measures
• Time-Series Models: Tag-based organization for efficient querying
• Correlation Schemas: Cross-module event correlation and anomaly detection
• Metadata Schemas: Asset, policy, dashboard, and user preference models
• API Models: Response formats, pagination, error handling, and streaming

Example

use llm_analytics_hub::schemas::events::{AnalyticsEvent, CommonEventFields, EventPayload};
use llm_analytics_hub::models::metrics::{MetricType, CounterMetric};
use chrono::Utc;

// Create a telemetry event
let event = AnalyticsEvent {
    common: CommonEventFields {
        event_id: uuid::Uuid::new_v4(),
        timestamp: Utc::now(),
        source_module: llm_analytics_hub::schemas::events::SourceModule::LlmObservatory,
        event_type: llm_analytics_hub::schemas::events::EventType::Telemetry,
        correlation_id: None,
        parent_event_id: None,
        schema_version: "1.0.0".to_string(),
        severity: llm_analytics_hub::schemas::events::Severity::Info,
        environment: "production".to_string(),
        tags: std::collections::HashMap::new(),
    },
    payload: EventPayload::Custom(llm_analytics_hub::schemas::events::CustomPayload {
        custom_type: "example".to_string(),
        data: serde_json::json!({"key": "value"}),
    }),
};

Re-exports

pub use database::Database;

pub use pipeline::ingestion::EventIngester;

pub use pipeline::ingestion::IngestionConfig;

pub use pipeline::ingestion::IngestionStats;

pub use analytics::anomaly::AnomalyDetector;

pub use analytics::anomaly::Anomaly;

pub use analytics::anomaly::AnomalyType;

pub use analytics::anomaly::AnomalySeverity;

pub use analytics::CorrelationEngine;

pub use analytics::AggregationEngine;

pub use schemas::events::AnalyticsEvent;

pub use schemas::events::CommonEventFields;

pub use schemas::events::EventPayload;

pub use schemas::events::EventType;

pub use schemas::events::Severity;

pub use schemas::events::SourceModule;

pub use models::metrics::AggregatedMetric;

pub use models::metrics::CounterMetric;

pub use models::metrics::GaugeMetric;

pub use models::metrics::HistogramMetric;

pub use models::metrics::MetricType;

pub use models::metrics::StatisticalMeasures;

pub use models::metrics::TimeWindow;

pub use models::timeseries::FieldSet;

pub use models::timeseries::IndexConfig;

pub use models::timeseries::RetentionPolicy;

pub use models::timeseries::TagSet;

pub use models::timeseries::TimeSeriesPoint;

pub use models::timeseries::TimeSeriesQuery;

pub use models::correlation::AnomalyCorrelation;

pub use models::correlation::CorrelationId;

pub use models::correlation::CorrelationType;

pub use models::correlation::EventCorrelation;

pub use models::correlation::EventGraph;

pub use models::correlation::RootCauseAnalysis;

pub use models::api::ApiError;

pub use models::api::ApiResponse;

pub use models::api::PaginatedResponse;

pub use models::api::PaginationMetadata;

pub use models::api::QueryResult;

pub use models::api::StreamEvent;

Modules

analytics

Analytics Engine

cli

CLI command implementations

common

Common utilities for CLI operations

database

Database Interaction Layer

infra

Infrastructure operations

models

Data models for metrics, time-series, correlation, and API responses

pipeline

Event Processing Pipeline

resilience

Resilience Patterns

schemas

Schema definitions for events and metadata

Constants

SCHEMA_VERSION

Schema version for data compatibility

VERSION

Version information