> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.nvidia.com/nemo/guardrails/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.nvidia.com/nemo/guardrails/_mcp/server.

# Observability Overview

> Logging, tracing, and metrics for end-to-end visibility into the behavior of the NVIDIA NeMo Guardrails library.

The NVIDIA NeMo Guardrails library exposes three observability signals so you can debug locally during development and monitor behavior in production: logs, traces, and metrics.
Each signal targets a different question and can be enabled independently of the others.

| Signal      | Best For                                                                                                                                                             | Page                              |
| ----------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------- | --------------------------------- |
| **Logging** | Debugging a single request with verbose console output, the `explain()` method, and the `log` generation option for structured per-request data.                     | [Logging](/observability/logging) |
| **Tracing** | Following a request through the rails it activated and the LLM calls it issued, with full OpenTelemetry semantic-convention support.                                 | [Tracing](/observability/tracing) |
| **Metrics** | Tracking aggregate behavior, including request volume, latency distributions, error rates, saturation, and per-LLM-call token usage for SLO dashboards and alerting. | [Metrics](/observability/metrics) |

Tracing and metrics use the OpenTelemetry library-instrumentation pattern: the NVIDIA NeMo Guardrails library depends on the OpenTelemetry API only, and the host application configures the SDK providers, exporters, and processors.