Mr. Dean! I'd like to express my gratitude for sharing such valuable content and information. Now, I have a question: Is it possible to delete and reprovision virtual machines using a scale set in a pool?
Thanks…that’s a Good question. Officially AVD does not support VM Scale Sets today. However they totally work. If you ran into issues with the scale set the AVD support team would not help…but the general Azure support would. And if the issue turns out to be the scale set interacting with AVD the answer will be to NOT use scale sets. If you understand that…you can do it at your own risk…the same thing is true for Ephemeral Disks.
Man... I can't help but feel like I like the Logic App approach more. You don't need to limit the maximum connections on the host pool and simply define, without percentages to figure out, when and how to power on additional hosts. The only thing I think I like here is the multiple scheduling plans along with the load balancing algorithm switching. Honestly - I will more than likely continue to use the Logic App approach.
Hi Dean. Great Video. Few questions: what is the time interval in which this runs or we can't mention any time interval? Is there a place where we can see the successful/failure run of this? Also while deploying it it seems it changed the host pool name from capital letters to small letters
I haven’t heard of it changing the letter case before…? The interval for checking the VMs is not currently published…the monitoring can be done like all other AVD resources, in diagnostic settings
@@AzureAcademy Thanks for the response. Let me know when the interval for checking is published. Also I have identified another BUG: while testing I found that it is enabling any session in drain mode by admins. There could be scenarios where an engineer may want to put the VM in drain mode to block the connection when they are fixing some issue on the particular WVD
I’m loving those learning videos. I need your guidance with a question. We’ve currently configured the start vm on connect feature to save cost. In addition to that So I’ve set a GPO to log off disconnected session after 1hr. Is there a way to shutdown personal VMs after a period of time.
You can have the Personal VMs power down on a schedule, I cover this in my Start VM on Connect video - ua-cam.com/video/GQeF77VVbmQ/v-deo.html However if you mean power down after the VM has been on for 3.5 hours, You can’t do that directly on Azure. But you could do it with Powershell or Azure Automation
Great feature. How to AVD will combine Start VM on connect with auto-scale feature ? Both methods will start VM based and their own threshold, isn't it ?
Azure has added autoscaling as an inbuilt custom role as Desktop virtualization power on and off role, please update and make a video on it, with 1 vm system on 24/7 active.
The scaling plan can only exist in certain regions. It's not available in all regions just yet. The whole schedule threshold logic is very confusing . I like the good old logic app approach as it takes the info from the log analytics and you can set the intervals. Not sure how the scaling plan gets that info and uses it for scaling.
I’m not sure what you are finding confusing about the schedule in the scaling plan…? You still set the time of day you want to handle more load and less load…and you have added flexibility of changing the load balancing algorithm…
Great job dude very easy to understand, just a quick one what if only one VM is running and it goes down for some reason bsod what ever, will that trigger a new VM to Start and allocate the users once they try to log back in? Many thanks
I don't think there ever will be...the logic all and Azure Automation account function outside of AVD so you can keep using it if you want...but it will cost $$$. The Autoscale feature is free. it is also native in the AVD service. I would recommend it
Hi Dean, I am missing the Enterprise applications, is there some way of registering these? I've tried the consent page but that appears to only register the old version of WVD apps, I've tried registering the Microsoft.DesktopVirtuilzation resource provider within Azure subscription to no avail. Any ideas? Thanks!
@@AzureAcademy Yeah, I thought the backend Enterprise App was still called Windows Virtual Desktop as when I labbed this a few months ago that was still the case, but looks like it's been added to Azure Virtual Desktop now :D
Hi Dean! Great video like every your video ❤️... Thanks to do that. I'm not sure if I understand correctly but if I put start on connect on my Pooled host pool, and I have an auto scale active... What happens? If I suppose 3 person per vm... And the first vm is full... When the fourth one try to connect the vm going on before the auto scale? Other question... With the logic app works every 15 minutes(by default) is here some fixed time from one check to the next one? Thanks
Thanks for watching! Using both the Autoscale plan and Start VM on Connect means that there is no waiting around. As soon as a user goes to log in a VM will be ready...or start up and be ready in a minute. Both cost nothing...so I say use both.
Reoccurrence has nothing to do with start be a man connect. It will only turn on a virtual machine if it is off when you try to connect. The Autoscale plan will turn VM‘s on and off if they’re not being used. If auto scale turns a VM off and then another user needs to login and there is no room on existing BM‘s start being on connect will turn another one on
@@AzureAcademy yes this is clear for me... Sorry maybe my question was not clear 😅. I talk only about the auto scale e not about start on connect ... Then just about the difference between logic app/automation and the new auto scale service.. With the logic app method we set up the recurring... I wish to know which is this recurring for the new method :)
The new Autoscale service will be more reactive then the logic app and automation account. When there are activities in the pool the Autoscale service Will evaluate what changes should be made
Does Autoscale continuously monitor Session Hosts and apply the rules, or does it run for example once a minute checking Session Host usage and applying the rules?
I think it's both schedule and trigger based. I've seen it run with 5 second intervals when there are changes in sessioncount. 11/4/2021 11:30:23 AM Active session hosts are at 23.44% capacity (15 sessions across 2 active session... 11/4/2021 11:30:28 AM Active session hosts are at 25.00% capacity (16 sessions across 2 active session... And every ~30 minutes when there's no changes in sessioncount: 11/4/2021 5:32:55 PM Active session hosts are at 3.12% capacity (1 sessions across 1 active session h... 11/4/2021 6:03:28 PM Active session hosts are at 3.12% capacity (1 sessions across 1 active session h...
Autoscale monitors for when there is a user session log in or log out in the host pool and categorizes these as session change events. When there are session change events, a scaling evaluation is triggered that will create logs. If there are no session change events or the events go down, Autoscale will check if there are any events that were missed. If there are several session change events that happen in a small time period, the scaling evaluations may be batched. Autoscale reacts relatively quickly to these session change events, but does not log on each one since that may cause overload.
Yes, I have with several customers. If it is not working for you…give more time before log off or open a support ticket to have your environment looked at
Mr. Dean! I'd like to express my gratitude for sharing such valuable content and information. Now, I have a question: Is it possible to delete and reprovision virtual machines using a scale set in a pool?
Thanks…that’s a Good question. Officially AVD does not support VM Scale Sets today. However they totally work. If you ran into issues with the scale set the AVD support team would not help…but the general Azure support would. And if the issue turns out to be the scale set interacting with AVD the answer will be to NOT use scale sets.
If you understand that…you can do it at your own risk…the same thing is true for Ephemeral Disks.
@@AzureAcademy I greatly appreciate your response to my inquiry.
👍👍
Man... I can't help but feel like I like the Logic App approach more. You don't need to limit the maximum connections on the host pool and simply define, without percentages to figure out, when and how to power on additional hosts. The only thing I think I like here is the multiple scheduling plans along with the load balancing algorithm switching. Honestly - I will more than likely continue to use the Logic App approach.
Thanks for the feedback!
Awesome content as always! Could you please share with me that group policies video? Thanks Dean!!
I've covered it a few video, this one is older, but the GPOs still apply
ua-cam.com/video/4zDazJsa2Zk/v-deo.html
Aother great video Dean...
Thanks for sharing.
Regards,
Sachin
Glad you enjoyed it
Hiya. Quick question does the Autoscale Scaling Plan load balancing algorithm setting (Breadth/Depth) override (take preference over) the Hostpool load balancing algorithm (Breadth/Depth)?
Yes it does. The host pool setting is updated by the scaling plan to force the load balancing algorithm you need at different times
Hi Dean. Great Video. Few questions: what is the time interval in which this runs or we can't mention any time interval? Is there a place where we can see the successful/failure run of this? Also while deploying it it seems it changed the host pool name from capital letters to small letters
I haven’t heard of it changing the letter case before…? The interval for checking the VMs is not currently published…the monitoring can be done like all other AVD resources, in diagnostic settings
@@AzureAcademy Thanks for the response. Let me know when the interval for checking is published. Also I have identified another BUG: while testing I found that it is enabling any session in drain mode by admins. There could be scenarios where an engineer may want to put the VM in drain mode to block the connection when they are fixing some issue on the particular WVD
The interval is everyone there is a change up the host pool. When session hosts power on or off, join or delete, change RDP properties etc.
Is it stopping things from working…? If not, I would not worry about it…if it is, open a support ticket
Using Nerdio for the Auto Scaling and image management.
Nerdio has a nice solution as well…the burst VMs are a big help as well
I’m loving those learning videos.
I need your guidance with a question. We’ve currently configured the start vm on connect feature to save cost. In addition to that So I’ve set a GPO to log off disconnected session after 1hr. Is there a way to shutdown personal VMs after a period of time.
You can have the Personal VMs power down on a schedule, I cover this in my Start VM on Connect video - ua-cam.com/video/GQeF77VVbmQ/v-deo.html
However if you mean power down after the VM has been on for 3.5 hours, You can’t do that directly on Azure. But you could do it with Powershell or Azure Automation
great explainations of the features Dean!
Thanks Charles!
Great feature.
How to AVD will combine Start VM on connect with auto-scale feature ?
Both methods will start VM based and their own threshold, isn't it ?
Start VM on Connect is t threshold based. It will power on a VM that is off when you go to connextb
It would be great if we can have a feature to add new hosts to existing hostpool (with same image as other hosts in that hostpool) using autoscale
You mean Autoscale should have the ability to create and destroy session hosts instead of powering them on and off…? I like it!
Azure has added autoscaling as an inbuilt custom role as Desktop virtualization power on and off role, please update and make a video on it, with 1 vm system on 24/7 active.
Yes it has…nice idea!
The scaling plan can only exist in certain regions. It's not available in all regions just yet. The whole schedule threshold logic is very confusing . I like the good old logic app approach as it takes the info from the log analytics and you can set the intervals. Not sure how the scaling plan gets that info and uses it for scaling.
I’m not sure what you are finding confusing about the schedule in the scaling plan…?
You still set the time of day you want to handle more load and less load…and you have added flexibility of changing the load balancing algorithm…
Great job dude very easy to understand, just a quick one what if only one VM is running and it goes down for some reason bsod what ever, will that trigger a new VM to Start and allocate the users once they try to log back in?
Many thanks
If a VM crashes then it will not be available for connections And other VMs will start to take the load
@@AzureAcademy Got My Host Pool in UK South so I can't use it.................
😔 I should be available soon
Hi Dean, I have an old style (LogicApp) scaling tool running. Any advice on how to decommission it properly before migrating to the new method?
I don't think there ever will be...the logic all and Azure Automation account function outside of AVD so you can keep using it if you want...but it will cost $$$. The Autoscale feature is free.
it is also native in the AVD service. I would recommend it
Hi Dean, I am missing the Enterprise applications, is there some way of registering these? I've tried the consent page but that appears to only register the old version of WVD apps, I've tried registering the Microsoft.DesktopVirtuilzation resource provider within Azure subscription to no avail. Any ideas? Thanks!
Don't worry, I have sorted this, should've searched via APP ID, rather than by name. It's been there all along!
Cool
I see
@@AzureAcademy Yeah, I thought the backend Enterprise App was still called Windows Virtual Desktop as when I labbed this a few months ago that was still the case, but looks like it's been added to Azure Virtual Desktop now :D
REALLY...cool I haven't seen that yet...THANKS!
Hi Dean! Great video like every your video ❤️... Thanks to do that. I'm not sure if I understand correctly but if I put start on connect on my Pooled host pool, and I have an auto scale active... What happens? If I suppose 3 person per vm... And the first vm is full... When the fourth one try to connect the vm going on before the auto scale? Other question... With the logic app works every 15 minutes(by default) is here some fixed time from one check to the next one? Thanks
Thanks for watching! Using both the Autoscale plan and Start VM on Connect means that there is no waiting around. As soon as a user goes to log in a VM will be ready...or start up and be ready in a minute.
Both cost nothing...so I say use both.
@@AzureAcademy thanks.... what about the recurrence for the new method?
Reoccurrence has nothing to do with start be a man connect. It will only turn on a virtual machine if it is off when you try to connect. The Autoscale plan will turn VM‘s on and off if they’re not being used.
If auto scale turns a VM off and then another user needs to login and there is no room on existing BM‘s start being on connect will turn another one on
@@AzureAcademy yes this is clear for me... Sorry maybe my question was not clear 😅. I talk only about the auto scale e not about start on connect ... Then just about the difference between logic app/automation and the new auto scale service.. With the logic app method we set up the recurring... I wish to know which is this recurring for the new method :)
The new Autoscale service will be more reactive then the logic app and automation account. When there are activities in the pool the Autoscale service Will evaluate what changes should be made
Hi..when will it be available in the Central Canada region
I can’t share details on that…just keep an eye on the supported regions section of the docs…won’t be long
Hi Antoine, just FYI, scaling plans are now available in Canada Central.
👍
Great as always!
Thanks…please share with others ☺️
Does Autoscale continuously monitor Session Hosts and apply the rules, or does it run for example once a minute checking Session Host usage and applying the rules?
Not sure of the exact interval, I know that it works fast
I think it's both schedule and trigger based.
I've seen it run with 5 second intervals when there are changes in sessioncount.
11/4/2021 11:30:23 AM Active session hosts are at 23.44% capacity (15 sessions across 2 active session...
11/4/2021 11:30:28 AM Active session hosts are at 25.00% capacity (16 sessions across 2 active session...
And every ~30 minutes when there's no changes in sessioncount:
11/4/2021 5:32:55 PM Active session hosts are at 3.12% capacity (1 sessions across 1 active session h...
11/4/2021 6:03:28 PM Active session hosts are at 3.12% capacity (1 sessions across 1 active session h...
Autoscale monitors for when there is a user session log in or log out in the host pool and categorizes these as session change events. When there are session change events, a scaling evaluation is triggered that will create logs. If there are no session change events or the events go down, Autoscale will check if there are any events that were missed. If there are several session change events that happen in a small time period, the scaling evaluations may be batched. Autoscale reacts relatively quickly to these session change events, but does not log on each one since that may cause overload.
Cool…
Thanks for the detail Seneca 👍
Has anyone actually got the notification message to work when forcing log off? The delay and log off notification seem to just not happen
Yes, I have with several customers. If it is not working for you…give more time before log off or open a support ticket to have your environment looked at
Awesome
Thanks mate
👍👍