To scale out the App Service plan automatically, you need to follow these steps:

1. Navigate to App Service and choose the App Service you want to scale out.
2. Click on Enable autoscale.
3. Then, you will need to configure the autoscale settings by specifying the following:
   • Autoscale setting name: Specify a descriptive name for the purpose of the autoscale
   • Resource group: Specify a resource group for the autoscale setting
   • Scale conditions: You can add scale conditions according to your needs by specifying the following:
     • Scale mode:
       • Scale based on a metric: Specify a metric based on which a scale out will be performed. For example, when CPU usage exceeds 70%, it will add one more instance.
       • Scale to a specific instance count: Specify the number of instances you want to scale to. If you have selected this option, you do not have to proceed with the upcoming settings.
     • Rules: Click on Add a rule to create a rule that will determine how the App Service plan will scale out/in. A new blade will open where you have to specify the following:
       • Metric source: Specify the source for which you will specify metrics that will determine the scale out/in.
       • Resource type: If you have selected Other resource in the metric source, you will have to specify the resource type.
       • Resource: If you have selected a metric source other than the current resource, you will need to specify that source. For example, if you selected Storage queue as a metric source, you will have to specify which storage account to use as a resource.
       • Time aggregation: This is the aggregation method that's used to aggregate sampled metrics. For example, time aggregation = average will aggregate the sampled metrics by taking the average of them.
       • Metric name: Specify which metric you want to measure so that the rule can determine whether to scale out/in.
       • Time grain statistic: This is the aggregation method within the timeGrain period. For example, statistic = average and timeGrain = PT1M means that the metrics will be aggregated every 1 minute by taking the average.
       • Operator: Specify the measure operator that will specify when the value of the metric has exceeded the threshold. For example, this operator will specify when the actual value of the resource is greater, lesser, equal to, and so on, compared to the threshold.
       • Threshold: This is the threshold on which the action will be performed to specify whether the plan scales out/in.
       • Duration: This is the duration of time required to look back for metrics. For example, 10 minutes means that every time autoscale runs, it will query metrics for the past 10 minutes. This allows the metrics to stabilize and avoids reacting to transient spikes.
       • Operation: Specify whether to decrease or increase instances count/percentage when a threshold is exceeded.
       • Instance count: The instance count will be increased/decreased according to the operation you specified in the previous step.
       • Cool down: The amount of time to wait after a scale operation before scaling again. For example, if the cool-down time is 10 minutes and a scale operation has just occurred, autoscale will not attempt to scale again until after 10 minutes. This is to allow the metrics to stabilize first. The following screenshot shows the scale rule:

• • • Instance limits: There are three instance limits that need to be specified
      • Minimum: Specifies the minimum instance count.
      • Maximum: Specifies the maximum instance count.
      • Default: If there's a problem reading the resource metrics and the current capacity is below the default capacity, then to ensure the availability of the resource, autoscale will scale out to the default. If the current capacity is already higher than the default capacity, autoscale will not scale in.
    • Schedule: By default, the first added rule is executed when none of the other scale conditions match. If you are adding other conditions, you can specify the start/end date or specific days for the rule:

4. Once you are done, click on Save.