escalating node count for autoscaling Azure app service

[Simon.Coles]

We have an application that has a mix of on-premise static infrastructure and newer services running in Azure.

We recently implemented a new .Net Core service as an Azure App Service with autoscaling.

The new service is appearing fine in our application flow map, but the node count is just increasing for every scale-up event, and not decreasing on scale-down.  Right now, it looks like we have 97 nodes, when in reality it's running at between 6-8 depending on load.

Is there a way to manage this automatically so the node count is correct for this service, preferably without static on-premise nodes disappearing through inactivity during quieter processing periods?

[Simon.Coles · Accepted solution]

We got this solved via a support ticket.

The node activity settings for retention period and deletion period are not available to SaaS customers.

However, the node reuse agent parameters are available for the .Net Microservices Agent, although this isn't in the documentation.

To set this up for our app running in Azure App Services, we had to add the following environment variables to the app service:

appdynamics.agent.reuse.nodeName=true
appdynamics.agent.reuse.nodeName.prefix=NodePrefix

where 'NodePrefix' is what you want to call the nodes (each instance will be suffixed with a number).

This needs to be set up for deployment slots as well as in the base app service config.

This is our set-up for our pre-production environment as an example.

Base app service configuration

Deployment slot