Auto Scaling provides the following benefits: automation, cost-effectiveness, high availability, flexibility, intelligence, and easy audit.

Automation

Auto Scaling performs automatic scaling based on your configurations to prevent the errors caused by manual intervention.

  • Scale-out activities:
    • Auto Scaling automatically creates the specified number of Elastic Compute Service (ECS) instances or elastic container instances and add them to your scaling group to provide sufficient computing power to meet your business requirements.
    • If your scaling group is associated with a Server Load Balancer (SLB) instance, Auto Scaling automatically attaches the ECS instances or elastic container instances in your scaling group to the SLB instance. This way, the SLB instance can forward requests to the attached ECS instances or elastic container instances.
    • If you set the Type parameter to ECS when you create your scaling group and associate an ApsaraDB RDS instance with your scaling group, Auto Scaling automatically adds the private IP addresses of the ECS instances in the scaling group to the whitelist that manages access to the ApsaraDB RDS instance. This way, application data on the ECS instances can be stored on the ApsaraDB RDS instance.
  • Scale-in activities:
    • Auto Scaling automatically removes the specified number of ECS instances or elastic container instances from your scaling group to release idle resources at the earliest opportunity.
    • If your scaling group is associated with an SLB instance, Auto Scaling automatically detaches the ECS instances or elastic container instances in the scaling group from the SLB instance. In this case, the SLB instance no longer forwards requests to the ECS instances or elastic container instances in your scaling group.
    • If you set the Type parameter to ECS when you create your scaling group and associate an ApsaraDB RDS instance with your scaling group, Auto Scaling automatically removes the private IP addresses of ECS instances in the scaling group from the whitelist that manages access to the ApsaraDB RDS instance. As a result, application data on the ECS instances can no longer be stored on the ApsaraDB RDS instance.

Cost-effectiveness

Auto Scaling provisions resources when the demand for the resources increases and releases resources when the demand for the resources decreases. This improves resource utilization and significantly reduces costs.

  • You do not need to prepare extra ECS instances or elastic container instances to ensure service availability during peak hours. You also do not need to worry about the waste of resources. Auto Scaling scales computing resources at the earliest opportunity. This significantly reduces your resource costs.
  • Auto Scaling also helps you save manpower and time.

High availability

Auto Scaling can detect whether an ECS instance or elastic container instance in your scaling group is healthy or whether the instance runs as expected. If Auto Scaling detects that an ECS instance or elastic container instance in your scaling group is not in the running state, it considers the instance unhealthy and replaces the instance with a new instance to ensure service availability.

Flexibility and intelligence

Auto Scaling provides a rich set of intelligent features that are suitable for diverse business scenarios and help simplify your configurations. This improves operational efficiency.

  • Scaling modes: Auto Scaling supports the fixed-number mode, health mode, scheduled mode, dynamic mode, and custom mode. You can combine these scaling modes based on your business requirements. In dynamic mode, Auto Scaling interconnects with the external monitoring system CloudMonitor by using API operations. For more information, see Scaling modes.
  • Instance configuration sources:
    • Auto Scaling supports more than one instance configuration source. For example, you can specify an existing instance as the instance configuration source or create a scaling configuration from scratch. If you set the Type parameter to ECS when you create a scaling group, you can also specify a launch template as the instance configuration source.
    • Auto Scaling also allows you to specify multiple instance types. This improves the flexibility of templates and increases the success rate of scale-out activities. For example, you can specify multiple instance types and disk categories in a template that is used to create ECS instances. You can also specify multiple vCPUs and memory sizes in a template that is used to create elastic container instances to determine the range of instance types. For more information, see Overview.
  • Scaling policies: If you set the Type parameter to ECS when you create a scaling group, Auto Scaling provides various scaling policies. If you set the Type parameter to ECI when you create a scaling group, the default scaling policy is used.
    • Priority policy: Auto Scaling preferentially scales instances in zones that have the highest priority. If the scaling activity fails, Auto Scaling scales instances in zones that have the next highest priority.
    • Balanced distribution policy: Auto Scaling evenly creates ECS instances in multiple zones to ensure high availability of the instances.
    • Cost optimization policy: If you specify multiple instance types in the instance configuration source, Auto Scaling preferentially creates ECS instances that have the lowest unit price of vCPUs and removes ECS instances that have the highest unit price of vCPUs.
    • Instance removal policy: You can remove ECS instances that are created from the earliest scaling configuration. You can also remove ECS instances that are created at the earliest or latest point in time.
    • Instance reclaim policy: You can release all ECS instances. You can also retain specific resources to reduce your costs.

    For more information, see Create a scaling group.

Easy audit

Auto Scaling records the details of each scaling activity. This way, you can effectively identify and troubleshoot issues.

Auto Scaling also provides the monitoring feature. You can use Auto Scaling together with CloudMonitor to monitor whether instances in your scaling group run as expected. This way, you can maintain service availability in an efficient manner.