We have recently announced the preview of a new feature in Windows Virtual Desktop allowing users to start their deallocated personal VM, more info is available in this blog post , and the announcment is here This allows users to power on the VM at just the time they need it. It also enables IT to reduce costs as much as possible as it is only powered on at the time the end user needs it. But how do you get the VM into a deallocated state in like manner, i.e. how can a user power off their VM? We do have an existing autoscaling tool, but that only supports pooled host pools today. We do have an "autoscaling service" under development which will bring support for personal host pools. There are other options available in Azure, such as each VM's Auto-Shutdown feature in VM section of the portal: However, this is set per VM, it also will just shut down the VM at the set time regardless of whether a user is still using this VM, or if they logged of at 6 PM meaning a waster ho
Showing posts from March, 2021
- Other Apps
Start VM on Connect entered public preview on March 31 2021 . Start VM on Connect does exactly what it says on the tin. It will start a deallocated (powered off from the hypervisor) or start a stopped (powered off from the OS) Azure VM, natively from within the WVD clients. The purpose besides the obvious of powering on the VM, is to achieve the maximum cost savings by keeping the VM deallocated, where the compute charges for the VM are not charged, for as much of the time as possible. Prior to this capability, IT could use the WVD Autoscaling tool , or Azure Automation to power on VM's at a time that IT predicts users will need it. So, for example, they may power on VM's at 08:30 for a 09:00 start. That however incurs 30 minutes of charges when the VM is powered on but not in use. However, what if the user is late or having a half day or is in fact not working at all that day. Then that increases those charges. Likewise, what if the user starts early, before the 08:30 start?