Auto Scaling is now available in all regions except edge-server regions. The following table details the use limits of this feature:
|One user under one region
- Up to 20 launch configurations can be created.
- Up to 20 scaling groups can be created.
- The maximum number of CVM instances that can be created depends on your CVM quota. For more information, see Purchase Limits for Pay-as-You-Go CVM Instances.
|A scaling group
- There can be only one launch configuration.
- Up to 2,000 CVM instances can be auto-scaled.
- Up to 100 scaling policies and 10 scheduled actions can be created.
- Up to 5 notifications can be created.
- Up to 10 lifecycle hooks can be created.
- The number of CVMs in all scaling groups cannot exceed the maximum number of IP addresses that the VPC subnet can provide.
- Currently, Auto Scaling does not support scaling up, which means it cannot automatically scale up the CPU, memory, or bandwidth of CVM instances.
- Auto Scaling and launch configurations are services supported at the region level. Therefore, you can only launch or terminate CVM instances in the same region.
- A scaling group and its associated CLB instances (in the case of a cross-region CLB instance, its backend VPC) must be in the same network environment (the VPC instance or the basic network in the same region).