Let AI Agents Own Your Monitoring: From Setup to Auto-Fixed Incidents

Checkly

Aug 20, 2026

Your AI agent can set up your monitoring, report on platform health, and open a PR to fix a failing check. All this works automatically.

In this webinar, Stefan (Developer Relations at Checkly, Google Developer Expert & Playwright Ambassador) demonstrates how Monitoring as Code makes Checkly agent-friendly by design, and walks through three agentic workflows live:

An agent builds your monitoring from scratch

Claude Code initializes Checkly in a Next.js project, discovers existing Playwright tests, converts the critical ones into production monitors, writes API checks and uptime monitors as typed constructs, validates everything with npx checkly test against real infrastructure, and deploys.

Agents report on your platform health

Using the Checkly MCP server, any agent tool can pull live monitoring data and deliver scheduled health briefs. There's no code access required. The agent queries your checks, test sessions, and failures directly from the Checkly infrastructure.

The highlight: the fully automated incident loop

A failing API check triggers a Claude Code routine via webhook. The agent checks out the source code, drills into the monitoring results, reads Checkly's AI root cause analysis, identifies the fix, validates it with a test session, opens a Linear ticket, and submits a pull request — before you've even opened your laptop.

Resources:
Checkly agent resources: https://www.checklyhq.com/docs/ai/overview/
Checkly CLI Docs: https://www.checklyhq.com/docs/cli/overview/
Checkly MCP Docs: https://www.checklyhq.com/docs/ai/mcp-server/

00:00 Welcome and agenda

00:49 Monitoring as code basics

01:58 CLI vs MCP for agents

04:16 Demo setup Next.js app

05:28 Initialize Checkly with AI

06:09 Skills built into CLI

07:46 Agent generates monitors

11:08 Test and deploy monitors

13:49 Add alert channels in code

15:40 MCP server reporting demo

18:17 Automated incident response loop