For a threshold-based autoscaling policy, this value is the initial number of instances to launch in the model deployment immediately after autoscaling is enabled. Note that anytime this value is updated, the number of instances will be reset to this value. After autoscaling retrieves performance metrics, the number of instances is automatically adjusted from this initial number to a number that is based on the limits that you set. Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.
For a threshold-based autoscaling policy, this value is the maximum number of instances the model deployment is allowed to increase to (scale out). Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.
For a threshold-based autoscaling policy, this value is the minimum number of instances the model deployment is allowed to decrease to (scale in). Note: Numbers greater than Number.MAX_SAFE_INTEGER will result in rounding issues.
The list of autoscaling policy rules.
Details for a threshold-based autoscaling policy to enable on the model deployment. In a threshold-based autoscaling policy, an autoscaling action is triggered when a performance metric meets or exceeds a threshold.