AgentTel FAQs

Question 1

What is AgentTel and what problem does it solve?

Accepted Answer

AgentTel enriches OpenTelemetry traces with structured, AI agent-ready context. It solves the problem of modern observability tools generating data primarily for human consumption, providing AI agents with the behavioral, topological, and decision metadata needed for autonomous incident diagnosis and resolution without human interpretation.

Question 2

What types of applications and environments does AgentTel support?

Accepted Answer

AgentTel provides full-stack observability for JVM backends (Java, Kotlin, Scala), browser frontends (TypeScript/JavaScript), and offers specialized GenAI Instrumentation for AI/ML workloads on the JVM, covering frameworks like Spring AI, LangChain4j, OpenAI, Anthropic, and AWS Bedrock SDKs.

Question 3

Is AgentTel easy to integrate into existing applications?

Accepted Answer

Yes, AgentTel is designed for easy integration. It offers a Spring Boot Starter for Spring Boot applications, a JavaAgent Extension for any JVM app (requiring no code changes), and a Web SDK for browser/SPA applications. Most enrichment is configured via YAML, simplifying setup.

Question 4

How does AgentTel enrich OpenTelemetry telemetry?

Accepted Answer

AgentTel enriches telemetry at three levels via YAML configuration: Topology (service identity, ownership, dependencies), Baselines (what 'normal' looks like for operations), and Decisions (what an AI agent is allowed to do, e.g., retryable, runbook URL). It also adds real-time Anomalies and SLO tracking.

Question 5

What is the Agent Interface Layer (MCP Server) and why is it important?

Accepted Answer

The MCP (Model Context Protocol) Server is a toolkit for AI agent interaction, exposing enriched telemetry as 'tools' AI agents can call. It provides health aggregation, structured incident context, a remediation framework, and action tracking, enabling agents to query live system state and execute actions.

AgentTel

AgentTel

主要功能

使用案例

主要功能

使用案例