Provide acceptable scale out time for instances in ASE Isolated instance types
Scaling is a core feature of App Services. We used ASE for the additional features however it means we are stuck with 40 mins scale out time to add an instance. Per support this is by design, however this is not documented anywhere in public domain.
Scaling is supposed to help us get customers on to Azure, if it takes 40 mins to add an instance, it will not meet the business needs, we abandoned ASE only due to this issue. Kindly document current scale out time in public documentation so that user know this before hand.
Ultimate goal should be to cut down scaling time to an acceptable level for ASE e.g. 2-5 mins to add instance as I understand ASE means dedicated instances.
Thanks for the feedback!
You’re right that ASE scaling operations take longer than the common plans on multi-tenant App Service. We are looking at ways to decrease the time so it will on par with multi-tenant but this means a new infrastructure is needed with managing a pool or pre-assigned VMs to the ASEs.
This is documented here, though without a time frame mentioned as it depends on a numbers of factors: https://docs.microsoft.com/en-us/azure/app-service/environment/using-an-ase#how-scale-works.
We will update when there is more information here.
Himanshu Trivedi commented
It would be nice to see ETA for scaling operation either on Azure Portal or REST API level.
With current options Engineers are helpless to provide any timescales to business or product owner. when performing ASE Front End Scale or even App Service Plan
+1 I would like to see improvement in provisioning times for plans within an ASE. In current state, we would have to set autoscale thresholds very low and still might not be scaled out in time before CPU / Memory pegs out.
Could you please update on this feature request? We are also on path of abandoning ASE for our Azure Function hosting due to horrible scale out time.