Skip to content

Autoscaling

Autoscaling adjusts resources using rules. Metrics, startup times and the limits of dependent services determine whether the application benefits.

Also known asAutomatic scaling
Cloud & Operations

What is autoscaling?

Autoscaling means automatically adjusting resources to an application’s needs. Rules can use metrics or schedules. Microsoft distinguishes increasing the capacity of individual resources from adding or removing instances. Which option is available automatically depends on the service used.

Example: additional portal instances

In this fictional example, a rule calls for additional portal instances when demand keeps growing. They first need time before they can serve requests. When load decreases, instances should be removed after their ongoing work has safely finished or been handed over.

The team chooses appropriate metrics and lower and upper limits. If the shared database already limits throughput, additional portal instances cannot automatically remove that limit. Their effect needs to be checked alongside the dependent services.

Scaling in is also part of the design. Faulty rules can cause unnecessary changes or insufficient capacity. Check both increasing and decreasing load, as well as the time until additional capacity becomes usable.

Sources

Articles on this topic

Cloud & Operations with ASES

Hosting, deployment, maintenance and support