Monitor AI workloads on Azure: Review

The first module ends with a question: what Azure service monitors AI workloads? The answer is Azure Monitor. That’s the level we're operating at here. This path tells you what the tools are named, shows you screenshots of their dashboards, and then tests you on their names. A real page-turner.

The section on Azure AI Foundry models sounds useful, talking about deployment and evaluation. But when it gets to the 'how,' the answer is just 'use Azure Monitor.' It’s a circular reference. I get that this is an introductory map. But as a builder, a map showing three towns named 'Monitoringville' isn't helpful. I need to know which roads are paved, what the speed limits are, and if the bridge is out. Instead, I got a series of links to product documentation.

The most frustrating part was on Log Analytics. It shows a picture of a KQL query but gives you neither the query itself nor a sandbox to run it in. This isn't a course; it's an interactive brochure. I did end up with a slightly better vocabulary for my next meeting with a client's DevOps team, which I guess is something. I can now confidently say, 'We should use Azure Monitor for that.' They will ask how. And I will send them a link to this learning path.