🔹 Feature: Azure Monitor Health Models
🔹 What It Does: Visually define and track the health of an entire workload — with each Azure resource’s signals rolled up through dependency relationships into a single, business-context-aware health state.
What Is It Giving You:
✅ State-Based Monitoring, Not Just Alert-Based: Assigns a health state to every managed entity — Healthy, Degraded, Unhealthy — instead of firing raw metric alerts. Answer “is the workload healthy?” not “which of the 200 alerts fired at 3am?”
✅ Alert Noise Reduction Through Correlation: Multiple signals per entity collapse into ONE health state. Multiple entities collapse into ONE workload health state. The 50-alert storm becomes one actionable signal.
✅ Dependency-Aware Rollup: Model relationships between components. When a database goes unhealthy, dependent services roll up as impacted. Root cause becomes visible in the graph — no more chasing symptoms across five different resource types.
✅ Graph View + Timeline View: Graph shows current health with dependencies at a glance. Timeline shows health over time — compare against SLOs, verify improvement after a fix, prove availability to the business.
✅ Reuses Everything You Already Collect: Platform metrics and resource logs, Prometheus metrics from Kubernetes clusters, VM logs and performance counters, Application Insights data. No new agents, no new data pipelines.
✅ Alert on Health State, Not Individual Signals: Fire alerts on unhealthy entities or on aggregate workload impact across multiple dependencies. Uses the same action groups you already have for existing Azure Monitor alerts.
✅ Graphical Designer With Auto-Discovery: Drag entities, define dependencies, add signals, tune rollup rules in a visual designer. Auto-populate the model from an Azure Resource Graph query, an Application Insights resource, or a service group — no manual entity-by-entity setup.
✅ Recommended Signals for Common Azure Resources: Start with Microsoft-defined health signals for common resource types (VMs, App Service, SQL, AKS, Storage), then customize with your own metric-based or log-query-based signals as you learn what “healthy” actually means for YOUR workload.
✅ SLO-Ready: Track availability of the workload and its components over time, compare against service level objectives, and prove business impact — not just infrastructure metrics.
🌐 https://learn.microsoft.com/en-us/azure/azure-monitor/health-models/overview
