-The Azure platform is designed to handle host node issues with minimal impact on VM performance. When a host node encounters a problem, Azure first attempts the least disruptive recovery method, which is to reboot the host. If rebooting the host isn't possible or if the issue is hardware-related, Azure initiates an automatic recovery action to take the faulty host out of rotation for further investigation. As part of this automatic recovery, a process known as service healing will automatically relocate all VMs on the faulty host to another healthy one. This process is usually completed within 15 minutes, although the recovery time can vary depending on factors such as the host's memory size and recovery methods used. Service healing is typically used as a last resort for hardware failures to ensure that VMs continue to operate without significant downtime.
0 commit comments