All Products
Search
Document Center

Container Service for Kubernetes:Configure a node pool by instance attributes

Last Updated:Mar 26, 2026

When you configure a node pool with specific instance types, scale-out can fail if those types run out of stock or go offline. To reduce this risk, configure the node pool by specifying instance attributes — such as vCPU count and memory — instead of fixed instance types. The node pool selects available instance types that match your criteria and automatically adds new matching types as they become available.

This feature is in a phased release. To enable it, submit a ticket.

Limitations

  • Review the ECS instance types not supported by ACK before configuring your node pool.

  • GPU instance types are not supported.

  • If the node pool billing method is Spot Instance, only the automatic bidding mode (SpotAsPriceGo) is supported. Setting a maximum price per instance (SpotWithPriceLimit) is not supported. For details, see Bidding modes.

  • vCPUs must be at least 4 and memory must be at least 4 GiB. For either resource, the Maximum value cannot exceed four times the Minimum value.

  • This feature cannot be used together with node autoscaling.

Configure a node pool by specifying instance attributes

Instance type selection priority follows the Scaling Policy configuration of the node pool. For details, see Create and manage a node pool.
  1. Log on to the ACK console. In the left navigation pane, click Clusters.

  2. On the Clusters page, click the name of the cluster you want to manage. In the left navigation pane, choose Nodes > Node Pools.

  3. Click Create Node Pool. In the Create Node Pool dialog box, complete the overall node pool configuration, then set Instance Configuration Mode to Specify Instance Attributes. The following table describes the configuration items specific to this mode. For all other configuration items, see Create and manage a node pool.

    If you edit a node pool to update the instance attribute configuration, the updated settings apply to new nodes only. Existing nodes in the pool are not affected.
    Configuration item Description
    Number of vCPUs The minimum and maximum number of vCPU cores. The maximum value cannot exceed four times the minimum value.
    Memory The minimum and maximum amount of memory. The maximum value cannot exceed four times the minimum value.
    Instance Family Level The performance tier of the instance family. Entry-level: lower cost, no guaranteed stable compute performance, suitable for workloads with low CPU utilization. Enterprise-level: suitable for enterprise scenarios with high business stability requirements. For a detailed comparison, see What are enterprise-level instances? What are entry-level instances? and What are the differences between enterprise-level and entry-level instances?.
    Instance Family Category The category of instance family, such as general-purpose, compute-optimized, or memory-optimized. For the full list, see Instance family classification and naming.
    CPU Architecture The CPU architecture: x86 or Arm.
    Instance Family The instance families to include in the selection range. For details, see Instance families.
    Instance types to exclude Instance types to exclude from scale-out. Use a wildcard character (*) to exclude a single instance type or an entire instance family. Examples: ecs.c6.large excludes that specific type; ecs.c6.* excludes all types in the c6 family.