All Products
Search

This Product

    ApsaraDB for MyBase:Configure resource overcommitment to reduce costs

    Document Center

    ApsaraDB for MyBase:Configure resource overcommitment to reduce costs

    Last Updated:Jun 20, 2026

    Overcommitment in ApsaraDB for MyBase lets you create multiple instances with the same specifications. This practice improves host CPU utilization and reduces resource costs. You can configure overcommitment both when you create a dedicated cluster and after it is created.

    Background

    In industries such as online gaming, predicting user growth for new servers can be difficult. Reserving dedicated ApsaraDB RDS servers may lead to wasted resources. In these cases, ApsaraDB for MyBase lets you overcommit resources to create multiple instances with the same specifications. You can then monitor host performance.
    • If user growth exceeds expectations, you can purchase additional hosts and migrate some instances to the new hosts.
    • If user growth is slower than expected, you can use overcommitment to meet service demands while significantly reducing resource costs.
    For example:
    • On a host with 32 cores and 256 GB of memory, you can create four 8-core, 64 GB instances without overcommitment. In this setup, resources are fully isolated.
    • On the same host, if you set the cpu overcommit ratio to 200%, you can create eight 8-core, 32 GB instances. You can add more hosts later as your business grows.

    Modify the overcommitment ratio

    1. Log on to the ApsaraDB for MyBase console.
    2. On the Dedicated Clusters page, find the target dedicated cluster, and in the Actions column, click EditEdit Dedicated Cluster.
    3. On the Edit Dedicated Cluster page, configure the CPU Overcommit Ratio, Storage Overcommit Ratio, and Maximum Memory Usage parameters as follows:
      • CPU Overcommit Ratio: The cpu overcommit ratio for the dedicated cluster. For example, a ratio of 200% allows the total CPU resources of all instances to be twice the physical CPU resources, which maximizes CPU utilization.
      • Storage Overcommit Ratio: The storage overcommit ratio for each host in the dedicated cluster.
      • Maximum Memory Usage: The maximum memory usage for each host in the dedicated cluster.
        Note You can modify the Maximum Memory Usage only for the MySQL engine.
      In the Edit Dedicated Cluster dialog box, you can also set the resource allocation policy (Balanced or Compact), the host failure handling policy (Auto-replace or Manual), and OS permission access (Enabled or Disabled). When you set the Maximum Memory Usage, reserve at least 10% of the host memory in the cluster to prevent out-of-memory (OOM) errors. After you complete the modifications, click OK.
    4. After clicking OK, find the cluster on the Dedicated Clusters page and click Details in the Actions column to view its configuration on the Basic Information page.
      Verify that both the CPU Overcommit Ratio and Storage Overcommit Ratio are updated to 200%.

    Scale out a dedicated cluster

    If host load becomes too high as your business grows, you can reduce it by adding new hosts or migrating instances. For instructions, see Add hosts and Migrate some instances to a new host.