Observability
Monitor and debug AI agents with OpenTelemetry-based distributed tracing and metrics
ADK-TS has built-in OpenTelemetry support. Every agent run, tool call, and LLM request automatically creates spans and records metrics — no manual instrumentation required. The data travels via OTLP to any compatible backend: Jaeger, Grafana, Datadog, New Relic, Honeycomb, or any OTLP endpoint.
How It Works
Initialize once before your first agent call. All subsequent agent, tool, and LLM operations emit spans and metrics automatically.
Quick Start
import { telemetryService, AgentBuilder } from "@iqai/adk";
await telemetryService.initialize({
appName: "my-agent-app",
otlpEndpoint: "http://localhost:4318/v1/traces",
appVersion: "1.0.0",
enableMetrics: true,
enableTracing: true,
});
const response = await AgentBuilder.withModel("gemini-2.5-flash").ask("Hello!");
console.log(response);
await telemetryService.shutdown(5000);No extra packages
All telemetry dependencies ship with @iqai/adk. Nothing extra to install.
Next Steps
🚀 Getting Started
Initialize and configure telemetry with all available options
🔍 Distributed Tracing
Understand what gets traced and how to create custom spans
📊 Metrics
Monitor performance with counters, histograms, and token tracking
🔌 Platform Integrations
Connect to Jaeger, Grafana, Datadog, and other OTLP backends
🔒 Production
Privacy controls, sampling, and best practices for production