Agent Observability

Monitor AI agent behavior with logs, traces, metrics, and cost telemetry. This skill covers the full observability stack for LLM-powered applications: from raw Prometheus counters to Grafana dashboards, OpenTelemetry tracing, structured logging, cost tracking, SLO definition, and PII redaction.

Examples use custom llm.* attributes, not a claim of current OpenTelemetry semantic-convention compliance. Import shared metric definitions once and use bounded configured label values. Telemetry export is data disclosure: approve destination/retention and omit prompts, credentials, direct user IDs and raw exception bodies. TLS is required for non-loopback collectors.

Core Metrics

Define these metrics at the application layer. All examples use the Prometheus client library naming conventions.

Latency

from prometheus_client import Histogram