Hey everyone - wanted to poll the crowd on an item. For those of you running in public cloud and utilizing Autoscale, how many of you have run into issues powering on after an off period? Example: Mass catalog power down in Azure (deallocated), then cannot bring the catalog back up (either partially or completely).

Hello George, on our end, we work with AWS. When we have failure linked to powering up resources, this is only due to capacity issue from AWS side. We are told to pay for capacity reservations to be sure they have sufficient capacity (lol). Other than that, it works perfectly.

There are a few challenges you can face depending on your architecture • zone availability. Not all zones are built equal. Assuming they are can cause chaos. Sometimes it’s smarter to not use zones at all for VDAs • The only way to guarantee your capacity is using capacity reservations, this is not the same as reserved instances and negates the use of Autoscale at all. Not all instances support reserved capacity • Citrix have some nice new capabilities in the form of backup instance sizes etc to help when your primary instance isn’t available Public cloud capacity kicks peoples butts all the time - Autoscale is great, but I’ve seen plenty of environments say no thanks, and just pay for the guarantee of their compute requirements

Yeah, what James says is true, it's annoying. We also submitted an RFE to Citrix to start resources in another availability zone in case the previous one is dry. If we need 20 machines for a delivery group, we over-provision in 2-3 AZ (20 in each) , it starts by splitting the load equally between AZs, may it fail to start in AZ1, it will try to start them in AZ2. But we say "do never start more than 20 machines in total". It's an on-going development so we hope it will come soon!

Thanks for all the feedback. I figured reserved capacity was going to be one of the answers. :)