This topic describes the limits of Auto Scaling.
Auto Scaling has the following limits:
- Applications deployed on the ECS instances in a scaling group must be stateless and horizontally scalable.
- ECS instances in a scaling group can be automatically released. Therefore, do not store application status information and related data such as sessions, databases, and logs on ECS instances. If applications deployed on these ECS instances require data to be saved, you can save the status information to independent ECS instances, databases to ApsaraDB for RDS, and logs to Log Service.
- Auto Scaling cannot add ECS instances to the whitelist of ApsaraDB for Memcache. You must manually add these instances to the whitelist.
- Auto Scaling can only scale the number of ECS instances and cannot scale vertically or adjust the vCPU, memory, and bandwidth specifications of ECS instances.
At present, the quantity limits of Auto Scaling are as follows:
- A single account can have up to 50 scaling groups in a region.
- A maximum of 10 scaling configurations can be created for a scaling group.
- A maximum of 50 scaling rules can be created for a scaling group.
- A maximum of six event notifications can be created for a scaling group.
- A maximum of six lifecycle hooks can be created for a scaling group.
- A maximum of five SLB instances can be associated with a scaling group at the same time.
- A maximum of 20 scheduled tasks can be created under an account.
- A maximum of 10 instance types can be specified in a scaling configuration.