Azure VM unavailable for 5 mins. RCA shows no issues found

A 5-minute VM availability alert can occur without a guest OS reboot. Azure documents two relevant patterns:

1. Platform-side interruption without guest reboot Host faults, auto-recovery, unplanned maintenance, or storage-related forced shutdowns can affect VM availability. In some cases Azure reboots the host or relocates the VM, and the guest OS may not show a normal reboot trail the way an in-guest restart would.
2. Availability reporting delay or inaccuracy during restart-like events Azure can briefly report VM availability inaccurately. During certain restart scenarios, monitoring can lag by 3 minutes 15 seconds, so short events may not appear consistently across Resource Health, Resource Graph, and Monitor.

Without a support plan, the practical self-service path is:

1. In the Azure portal, open the affected VM.
2. Check Help > Resource health.
   • Review the health event around 03:03–03:08 on 7 August 2026.
   • Azure may add root-cause details there up to 72 hours after the initial unavailability.
3. Run VM diagnostics:
   • Go to Diagnose and solve problems > Common problems > VM restarted or stopped unexpectedly.
   • In Tell us more about the problem you are experiencing, select My resource has been stopped unexpectedly.
   • Review the diagnostics output for RCA details.
4. Correlate with Azure Monitor alert timing:
   • For VM availability alerts, Azure recommends using the VM availability metric.
   • If the event was only 5 minutes, compare the alert fire time with Resource Health and diagnostics results to determine whether it was a short platform event versus a monitoring/reporting delay.
5. Check for guest evidence only as confirmation:
   • Since Ubuntu shows no reboot and no heartbeat issue, that supports the possibility of a short platform-side availability event rather than an in-guest crash.
6. If this repeats, consider mitigation:
   • For production workloads, place two or more VMs in an availability set to reduce impact from planned or unplanned maintenance.
   • If repeated host-related issues affect the same VM, Azure notes that manually redeploying the VM to another host can help. This changes the underlying host, so use it carefully and only during a maintenance window.

If Resource Health and diagnostics remain empty after 72 hours, the available self-service evidence is limited to the alert timeline, Resource Health history, and diagnostics output. The documented no-support-plan option is to use those portal diagnostics rather than opening a support request.

References:

• How to check Root Cause Analysis of unexpected Azure virtual machine reboots
• Understand a system reboot for Azure VM
• Monitor virtual machines with Azure Monitor: Alerts
• VM availability might be inaccurate during VM restarts