Your Agent Shipped the Fix. Did It Work?

Coding agents made writing and shipping code fast. Knowing whether what shipped works in production is still manual, and most of the tooling for it assumes an SRE or DevOps team you may not have. This hands-on workshop covers the other half of the loop. How do you give a coding agent enough production context to diagnose a real failure, and how do you verify that the fix it proposed reached production and made things better? You get your own GitHub Codespace running a deliberately broken microservice app under live traffic. We work through detection, causal root cause, a pull request, a deploy, and verification against error rates and SLOs, with visible proof in the running app. Then you point the same loop at a repository of your own. We use Bluebox by Dynatrace, a free helper agent, to run the loop, and leverage OpenTelemetry and your code repo as data sources. Bring a laptop, a GitHub account, and a repo where you can install a GitHub App. We will provide the microservice app, coding agents and everything else runs in the browser.