This topic describes Auto Scaling resources and their usage to help you understand the core concepts and features of Auto Scaling.
Usage process
This topic describes how to use Auto Scaling in the Alibaba Cloud Management Console. For information about how to use Auto Scaling by calling API operations, see Process.
If you set the Instance Configuration Source parameter to Launch Templates or Select Existing Instance when you create a scaling group, a scaling configuration is automatically created and becomes active. You do not need to manually recreate a scaling configuration. After the process for creating a scaling group is complete, you can directly enable the scaling group.
Create a scaling group
A scaling group consists of identical instances designed for use in similar business scenarios. When creating a scaling group, you can define attributes such as the minimum and maximum instance limits, the template for scale-out events, and the scale-in policies. For more information, see Overview.
Create a scaling configuration
A scaling configuration is a template used by Auto Scaling to automatically launch Elastic Compute Service (ECS) instances or elastic container instances. For more information, see Overview.
Enable the scaling configuration
If a scaling group lacks an active scaling configuration, a prompt will appear asking you to enable a scaling configuration after the scaling group is created. A scaling group can have multiple scaling configurations, but only one can be active at any given time. For more information, see Manage scaling configurations.
Enable the scaling group
Scaling events occur only in scaling groups that are in the Enabled state. When a scaling group has an active scaling configuration, you will be prompted to enable the group. You can also go to the Scaling Groups page to manually enable a scaling group. For more information, see Enable a scaling group.
Creates a scaling rule
A scaling rule defines the action for adjusting the number of instances in a scaling group, helping manage the addition or removal of instances. For more information, see Overview.
Execute the scaling rule
To execute the scaling rule, you can use one of the following methods. For more information, see Execute a scaling rule.
Manual execution: Manually execute a scaling rule. This method is ideal for handling temporary business needs.
Automatic execution based on scheduled tasks: Configure a scheduled task to execute a scaling rule at a specified time. This method is ideal for managing workload fluctuations with predictable patterns.
Automatic execution based on event-triggered tasks: Configure an event-triggered task to monitor specific metrics, report alerts, and execute a scaling rule. This approach is ideal for handling workload fluctuations without predictable patterns.
Tutorial
Scenarios
Scenario description and example | References |
For temporary business needs, manual adjustment of servers is required because determining the number of servers to add and the timing for adding the servers is challenging, especially when no monitoring metric is available. For example, if a company requires additional servers for testing purposes due to temporary business needs, manual oversight is required to determine the number of servers to add or remove and the timing for these adjustments. | |
If your workload fluctuations follow a predictable pattern and distinct peak and off-peak hours can be identified based on time, you can use Auto Scaling to automatically increase or decrease the number of ECS instances at designated time points accordingly. For example, a game company experiences a consistent increase in business workloads from 18:00:00 to 23:00:00 nightly, requiring the addition of servers during this period. However, only a fixed number of servers are required to maintain regular operations during other periods. | |
If your workload experiences dynamic fluctuations, you may find it challenging to determine the optimal number of servers and the timing for adding or removing servers. In this case, if you intend to adjust the server capacity based on real-time workloads, you can create event-triggered tasks in the Auto Scaling console. For example, the traffic of a news website is unpredictable. A surge of page views occurs with breaking news, but the views drop when the news loses its timeliness. |