Inside the Observability Agent: How Deep Investigations and Reasoning Work (opens in new tab)

Incidents are rarely explained by one chart. A spike in request failures might be caused by a dependency timeout. A dependency timeout might come from a connection pool issue. And that connection pool issue might trace back to a deployment or a resource health event that started minutes earlier. From the outside, Azure Copilot observability agent presents this as a focused troubleshooting experience that correlates signals such as metrics, logs, alerts (and alert context), tracing data, and r...