You can create multiple instances of the same specifications in an ApsaraDB for MyBase dedicated cluster by configuring resource overcommitment. This improves CPU utilization
of the hosts and reduces costs. This topic shows you how to configure resource overcommitment
for dedicated clusters when or after you create the clusters based on your needs.
Background information
In industries such as the game industry, the user growth trend of a new service is
unpredictable. If you purchase ApsaraDB RDS reserved instances, the instance resources
may be wasted. The resource overcommitment feature allows you to create multiple instances
of the same specifications on your host and monitor the host performance.
- If the number of users grows faster than expected, you can purchase additional hosts
and migrate some instances to the new hosts.
- If the number of users grows slower than expected, you can configure resource overcommitment
to meet business requirements and reduce resource costs.
Example:
- If you do not configure resource overcommitment, after you purchase one 32-core 256
GB host, you can create four 8-core 64 GB instances on the host. The resources of
these instances are isolated from each other.
- If you set the CPU overcommitment ratio to 200%, after you purchase one 32-core 256
GB host, you can create eight 8-core 32 GB instances. If the number of users grows
fast, you can add hosts later.
Set overcommitment ratios when you create a dedicated cluster
- Log on to the ApsaraDB for MyBase console.
- In the upper-left corner of the RDS Dedicated Clusters page, select the region where you want to create the dedicated cluster.
- On the RDS Dedicated Clusters page, click Create Dedicated Cluster.
- In the Create Dedicated Cluster panel, set CPU Overcommit Ratio and Storage Overcommit Ratio. The following content describes the parameters:
- CPU Overcommit Ratio: the CPU overcommitment ratio of the dedicated cluster. The CPU overcommitment ratio
ranges from 100% to 300%. The default ratio is 200%. If you use the default setting,
the CPU resources that can be allocated are twice as many as the CPU resources that
are provided. This maximizes CPU utilization.
- Storage Overcommit Ratio: the storage overcommitment ratio of each host in the dedicated cluster. The storage
overcommitment ratio ranges from 100% to 200%. The default ratio is 100%.

- Click Create. On the RDS Dedicated Clusters page, view the ApsaraDB for MyBase dedicated cluster that you created.
Set overcommitment ratios after you create a dedicated cluster
After you create a dedicated cluster, you can set overcommitment ratios based on the
host specification and the instance specification. The following content provides
examples:
- If your host specification is 4 GB memory per core and your instance specification
is 2 GB memory per core, you can set the CPU overcommitment ratio to 100%.
- If your host specification is 8 GB memory per core and your instance specification
is 2 GB memory per core, you can set the CPU overcommitment ratio to 300%.
- Log on to the ApsaraDB for MyBase console.
- On the RDS Dedicated Clusters page, find the dedicated cluster for which you want to set overcommitment ratios
and click Edit in the Actions column. The Edit Dedicated Cluster panel appears.
- In the Edit Dedicated Cluster panel, set CPU Overcommit Ratio and Storage Overcommit Ratio. The following content describes the parameters:
- CPU Overcommit Ratio: the CPU overcommitment ratio of the dedicated cluster. The CPU overcommitment ratio
ranges from 100% to 300%. The default ratio is 200%. If you use the default setting,
the CPU resources that can be allocated are twice as many as the CPU resources that
are provided. This maximizes CPU utilization.
- Storage Overcommit Ratio: the storage overcommitment ratio of each host in the dedicated cluster. The storage
overcommitment ratio ranges from 100% to 200%. The default ratio is 100%.

- Click OK. On the RDS Dedicated Clusters page, find the dedicated cluster and click Details in the Actions column. On the Basic Information page, view the configuration information.
Scale out a dedicated cluster
To reduce workloads on your hosts, add hosts in the dedicated cluster and migrate
some instances to the new hosts. For more information about how to add hosts and migrate
instances, see Create hosts and Migrate an instance across hosts.