When you need more processing and memory power, use horizontal scaling and scale out to increase your Azure Data Explorer cluster to the required size and for the necessary time frame. Once the load is reduced, you can scale in the cluster to its original size.
This scaling is performed by defining the precise number of instances you want to use with Manual scale. Alternatively, you can create your own rules based on the available metrics using Custom autoscale. Now, you can use Optimized autoscale to allow Azure Data Explorer to control the size of the cluster and optimize the performance and cost. To ascertain that you don’t exceed the planned budget, you can define the maximum cluster size.
Optimized autoscale is the recommended way to manage your Azure Data Explorer cluster size. It uses the service performance counters to make sure that you have enough resources when you need them. When the load decreases, the Optimized autoscale reduces the scale of the cluster to minimize costs.