Set up monitoring for CPU and Storage in Azure

Hello Thea, you’re on the right track—Azure Monitor can cover both CPU and storage for Managed Instances, and you can use a mix of metric alerts and log-search alerts to get exactly the thresholds and notifications you want. Here’s an outline of how to approach each ask, plus how to test alerts and get direct Microsoft support:

1. Alerting on High CPU for Managed Instances • Use a metric alert on the built-in “CPU percent” metric – In the Azure portal, open your Managed Instance → Alerts → + New alert rule – Condition → Select Metric → search for “CPU percent” → Aggregation (e.g. Average over 5 min) – Threshold type: Static, e.g. > 80% → Set frequency (evaluation window) – Attach an Action Group (email/SMS/Logic App/etc.) • Best practice: pick a threshold you’ll have time to react on (e.g. 80% for 5 min or 90% for 1 min) and suppress noise with a short “auto-resolve” window.
2. Alerting on Low Storage for Managed Instances Option A: Native metric, if exposed – Some PaaS services expose a “Storage percent used” or “Available storage” metric. You can follow the same steps as above but select that metric instead of CPU. Option B: Log-search alert via Log Analytics – First, configure Diagnostic Settings on your MI to send SQL Insights metrics to a Log Analytics workspace – In that workspace, create a log query like (adapt to your resource names):

        InsightsMetrics  
      
      
        | where Resource == "your-mi-name" and Namespace == "LogicalDisk" and Name == "FreeSpacePercentage"  
      
      
        | summarize arg_max(TimeGenerated, *) by Computer, Disk  
      
      
        | where Val < 10   // e.g. less than 10% free  
   

   – Save that query and click “New alert rule” from the Logs blade – Set the alert logic: whenever the result count > 0 within the evaluation period – Tie in an Action Group as above • Pro tip: you can parameterize the KQL threshold (here “Val < 10”) or leave it flexible and enforce thresholds in the alert rule UI.
3. Testing Your Alerts • In the Azure portal, on an existing alert rule, there’s a “Test” button under the Action Group section—use that to fire a sample notification • For metric alerts you can temporarily set absurdly low thresholds (e.g. CPU > 1%) to force an alert, then revert when you’ve confirmed delivery • For log alerts, you can tweak your KQL to return one dummy row (e.g. union print Val=0) and confirm the rule fires
4. Templates & Best-Practice Guides • Common VM-based storage alerts: “Best practices for VM monitoring using Azure Monitor alerts” covers disk and CPU counters you can adapt for PaaS workloads • The “Monitoring Disk Space with Azure Monitor for Virtual Machines” article shows KQL patterns you can reuse • Azure Monitor Alerts overview: general concepts around static vs. dynamic thresholds, aggregation types and action groups
5. Getting a Short Microsoft Meeting If you want a live, interactive run-through with a Microsoft engineer: • Open a support request in the Azure portal (Help + support → New support request) • Choose “Service and subscription limits (quotas)” or “Configuration and management”—then you’ll see an option for “Architectural guidance” or “Proactive consultation” (depends on your support plan) • From there you can request a 30- or 60-minute call to walk through your Managed Instance alerting design

Hope this helps!

Reference List

1. Monitor disk space with Azure Monitor for VMs (KQL & portal) https://docs.microsoft.com/azure/azure-monitor/vm/vm-insights-disk-space
2. Best practices for VM monitoring using Azure Monitor alerts https://docs.microsoft.com/azure/azure-monitor/vm/monitor-virtual-machine-alerts
3. Azure Monitor Alerts overview https://docs.microsoft.com/azure/azure-monitor/platform/alerts-overview
4. Monitor Azure SQL Database (metrics vs logs storage) https://docs.microsoft.com/azure/azure-sql/database/monitoring-sql-database-azure-monitor