Set compute auto-scaling

This topic describes how to enable compute auto-scaling for a cluster instance. After this feature is enabled, when the CPU usage of any node exceeds the specified threshold, a scale-up is automatically triggered. Each time a scale-up is triggered, compute resources (CPU and memory) are doubled synchronously until the maximum specification you set is reached.

Background information

If your business involves a large amount of data or your business volume fluctuates frequently, we recommend that you enable compute auto-scaling. When CPU resources cannot meet the performance requirements during business peaks, the cluster CPU can be automatically scaled up to ensure the stability of your online business.

• CPU auto-scaling applies to the CPU resources of full-featured replica nodes.

• The system checks node CPU usage on a second-by-second basis. When the CPU usage is greater than or equal to the specified value, a scale-up is automatically triggered. You can enable or disable this feature at any time after the instance is created.

Prerequisites

Before you set compute auto-scaling, make sure that the following conditions are met:

• Only project owners and project admins can perform this operation.

• The cloud service provider of the current cluster is not AWS.

• The storage architecture of the current cluster is Shared Nothing.

• The storage type of the current cluster is cloud disk.

Considerations

• After you enable compute auto-scaling, second-level monitoring for the instance is enabled synchronously.

• Second-level flash disconnections may occur during scale-up. Make sure that your application is configured for automatic reconnection.

• We recommend that you disable tenant resource isolation to ensure that scale-up takes effect immediately. We also recommend that you configure alerts for auto configuration change failures to avoid business impact due to insufficient resources.

Procedure

1. Log in to the OceanBase Cloud console.

2. In the left-side navigation pane, click Instances.

3. In the instance list, find the target cluster, and click the cluster instance name to go to the Overview page.

4. On the Overview page, click ... in the upper-right corner and select Set Compute Auto-scaling; or on the right side of the topology section, click Auto Scale-Up > Set Compute Auto-scaling.

5. In the dialog box that appears, enable the compute auto-scaling switch, configure the following parameters, and click OK.

   Parameter	Parameter Description
   CPU Utilization	The CPU usage of nodes. During normal hours, when the CPU usage of a node reaches this preset threshold, compute auto-scaling is triggered. The default value is 70%. Valid values range from 10% to 100%.
   Observation window	The continuous monitoring duration before auto-scaling is triggered. When the CPU usage of any node in the instance reaches the threshold, the system enters the observation state. If the percentage of time during which the CPU usage exceeds the threshold within the specified duration reaches the specified value, a scale-up is performed.
   Threshold Breach Duration %	The percentage of sampling points during the observation window during which the CPU usage exceeds the threshold. When this percentage reaches the specified value, a scale-up is triggered. The default value is 80%. Valid values range from 10% to 100%.
   Max CPU Scale-Up Limit	Set the auto-scaling upper limit based on your business needs. If CPU auto-scaling reaches the upper limit you set, auto-scaling fails.

   Note

   After compute auto-scaling is set, it will be triggered under specific conditions, and the instance will be marked with "Auto" in the instance list.