lexigram-ai-observability

AI observability for the Lexigram Framework — tracing, metrics, and monitoring

Overview

AI-layer observability for the Lexigram Framework. Provides tracing, metrics, health monitoring, and decorator-based instrumentation for LLM calls, RAG operations, and vector store interactions — all wired through the DI container via ObservabilityModule. Zero-config usage starts with sensible defaults.

Full documentation: docs.lexigram.dev

Install

uv add lexigram-ai-observability
# Optional extras
uv add "lexigram-ai-observability[opentelemetry]"

Quick Start

from lexigram import Application
from lexigram.di.module import Module, module

from lexigram.ai.observability import ObservabilityModule
from lexigram.ai.observability.config import ObservabilityConfig

@module(imports=[
    ObservabilityModule.configure(
        ObservabilityConfig(
            enabled=True,
            metrics_enabled=True,
            tracing_enabled=True,
            health_checks_enabled=True,
        )
    )
])
class AppModule(Module):
    pass

app = Application(modules=[AppModule])
if __name__ == "__main__":
    app.run()

Configuration

Zero-config usage: Call ObservabilityModule.configure() with no arguments to use defaults.

Option 1 — YAML file

# application.yaml
ai_observability:
  enabled: true
  metrics_enabled: true
  tracing_enabled: true
  health_checks_enabled: true

Option 2 — Profiles + Environment Variables (recommended)

export LEX_AI_OBSERVABILITY__ENABLED=true
# Environment variables for each field

Option 3 — Python

from lexigram.ai.observability.config import ObservabilityConfig
from lexigram.ai.observability import ObservabilityModule

config = ObservabilityConfig(
    enabled=True,
    metrics_enabled=True,
    tracing_enabled=True,
    health_checks_enabled=True,
)
ObservabilityModule.configure(config)

Config reference

Module Factory Methods

Key Features

• Tracing: Distributed tracing for LLM calls, RAG pipeline stages, and vector store queries
• Metrics: Token usage, latency, error rates, and cache hit ratios
• Health monitoring: Background health checks for AI components
• Decorators: @trace_llm, @trace_rag, @track_llm_call for automatic instrumentation
• Observable wrappers: ObservableLLMClient and ObservableVectorStore
• OpenTelemetry support: Optional OpenTelemetry integration

Testing

async with Application.boot(modules=[ObservabilityModule.stub()]) as app:
    # your test code
    ...

Key Source Files