CODEWITHSUNDEEP
azureazure-app-service-plans
Youngjae
Youngjae

Reputation: 25080

Azure App Service Architecture understanding: Per-App Scaling

Regarding this MSDN article; https://msdn.microsoft.com/en-us/magazine/mt793270

Per-App Scaling section has below paragraphs.

Using the earlier 50 app example, with per-app scaling enabled for the App Service Plan, all 50 apps can be assigned to the same App Service Plan. Then, the scaling characteristics of individual apps can be modified:

  • 40 low-volume applications set to run on a maximum of a single server each.
  • Five mid- to low-volume applications set to run on a maximum of two servers each.
  • Five remaining high-volume applications set to run on a maximum of 10 servers.

The underlying App Service Plan can start out with a minimum of five servers. And then auto-scale rules can be set to scale out as needed based on memory pressure vs. CPU.

I cannot understand why above three configuration gives minimum of five servers as a starting count.

Does it say 40 low-volume apps to 1 server, 5 mid-volume apps to 1 servers, and rest 5 high-volume apps to (undescribed) 3 servers?

Upvotes: 2

Views: 1729

Answers (2)

Youngjae
Youngjae

Reputation: 25080

In my understanding, a minimum of five servers statement is calculated as below setup.

  • 40 small apps are set to 1 server, fixed value.
  • 5 medium apps are set to minimum 1 to maximum 2 server for each.
  • 5 large apps are set to minimum 1 to maximum to 10 server for each.

Therefore, "The underlying App Service Plan can start out with a minimum of five servers" are given by the number of large apps. Each large app are allocated to each server and rest of them (40small+5mid) might be belonged to one of five.

Please leave a comment if my understanding is not correct.

Upvotes: 0

David Makogon
David Makogon

Reputation: 71118

That article is giving examples of splitting up your numerous apps across multiple app services.

The TL;DR version: With an App Service plan, every app deployed to that app service plan resides on every instance. If you have an App Service plan with 3 instances, then every one of your apps lives on those 3 instances. Go to 4 instances? Again - all your web apps, api apps, etc. go onto all 4 instances.

If that scale model doesn't work for you (e.g. you have a super-high-load app that eats up all the memory or all the cpu), then maybe that individual app is a candidate for its own app service plan, so that it doesn't starve the other apps in the app service plan.

That's really it. Want to put all your apps in a single app service plan? Great - do it. You have a high-demand app causing issues with your low-volume apps? Great - move it to its own app service plan. Totally up to you.

But that article... is just giving an example of splitting up apps into several app service plans.

Upvotes: 3

Related Questions