All Products
Search
Document Center

Container Service for Kubernetes:Smart Hosting Mode Overview

Last Updated:Feb 28, 2026

Smart Hosting (Auto Mode) provides automated and intelligent cluster O&M features for enterprise production environments. When enabled, you can create a Kubernetes cluster that follows best practices with simple network planning configurations and leverage automated cluster O&M capabilities.

With Auto Mode, you gain:

  • Comprehensive managed O&M, including fully managed cluster control plane and critical system components, automatic Kubernetes version upgrades, and full lifecycle management for nodes.

  • Automatic node scaling, eliminating manual capacity planning and optimizing node resource costs.

  • Default best security practices, which include immutable node host file systems to strengthen system security protection.

image

Auto Mode is applicable to general container scenarios, such as the following:

  • If you have limited resources and are unable to dedicate significant human resources to Kubernetes infrastructure management.

  • If you lack professional Kubernetes O&M experience but want to containerize and modernize applications using Kubernetes.

  • To quickly deploy standardized Kubernetes environments globally.

Features

  • Managed O&M

    • Cluster level: The cluster control plane and critical system components (such as CoreDNS and CSI) are fully managed. Automatic Kubernetes version upgrades are enabled by default.

    • Node level: By default, an Auto Mode node pool (referred to as an Auto Mode node pool) is created. It automatically recommends optimal instance types, which eliminates manual configuration. ACK also handles O&M responsibilities, such as operating system version and node component upgrades, and node security vulnerability fixes.

  • Node Elastic Scaling

    • Built-in node instant elasticity dynamically scales nodes based on workload demands. This eliminates the need for prior capacity planning and optimizes resource costs.

  • Security Protection

    • Optimized base software stack by default. Nodes use ContainerOS. Its immutable root file system strengthens security protection. The streamlined system and configuration accelerate node startup. An optimized kernel fully leverages hardware resource performance.

    • Security and observability features are enabled by default, such as Cluster inspection, API Server auditing, and control plane component log collection. These features help quickly identify and resolve cluster issues.

Important
  • When ACK performs O&M responsibilities, such as operating system version upgrades, software version upgrades, or security vulnerability fixes, operations such as software version upgrades, software configuration modifications, restarts, and node draining are involved. Avoid manual O&M on ECS nodes in the node pool, such as restarting, data disk mounting, or logging in to modify node configurations. This prevents conflicts with automation policies.

    Also, set the number of replicas for workloads, PreStop graceful shutdown policies, and PodDisruptionBudget policies appropriately. This ensures nodes can be drained for O&M without service interruption.

  • Because nodes may be automatically replaced by elastic scaling policies, avoid storage methods that rely on specific node paths, such as HostPath. Use PersistentVolumeClaims (PVCs) for persistent storage.

  • ContainerOS has an immutable root file system. If your application requires a customized node environment or local persistent data, see ContainerOS to assess its compatibility.

Get Started with Smart Hosting Mode

References

Shared Responsibility Model

The ACK managed cluster Smart Hosting Mode aims to provide automated and intelligent Kubernetes cluster O&M features, reducing your effort in Kubernetes cluster O&M. However, in some scenarios, you still have certain responsibilities.

Alibaba Cloud is Responsible

Customer is Responsible

Shared Responsibility

  • Deployment, maintenance, and upgrade of the cluster control plane.

  • Installation, configuration, and upgrade of core cluster components.

  • Automatic scale-out and scale-in of node pools, operating system upgrades, software version upgrades (including CVE security vulnerability fixes), and more.

  • Basic cluster information configuration, such as network planning and VPC configuration.

  • Setting and managing cluster RAM permissions and RBAC.

  • Deployment, O&M, and proper configuration of application workloads. Proper configuration includes setting the number of replicas, PreStop graceful shutdown policies, and PodDisruptionBudget policies to ensure nodes can be drained for O&M without service interruption.

  • Promptly receive monitoring and alerting for clusters and applications, and respond based on alert information.

  • Overall cluster security. Cluster security responsibilities apply to the shared responsibility model. For more information, see Shared Responsibility Model for Security.

  • Troubleshooting and problem resolution.

Billing Description

Auto Mode is only available in ACK Managed Cluster Pro Edition. The feature itself is free. Cluster management fees and related cloud product fees are charged as usual. For more information, see Billing overview and Cloud Product Resource Fees.

Quotas and Limits

  • Auto Mode node pools dynamically scale nodes based on workload demands. By default, they support scaling out to a maximum of 200 nodes. You can modify the maximum number of instances using the node pool scaling feature.

  • Auto Mode node pools do not support instance types such as Arm or local disks. They only support ContainerOS 3.6 and later. ACK's recommended default instance types meet application requirements for most scenarios. You can adjust them as needed in the console. Set a sufficient number of instance types to enhance the node pool's elasticity and prevent scaling failures.

  • If the cluster is large or the account has many resources, follow the quotas and limits for ACK clusters.

    • ACK configuration limits (such as account balance) and single-cluster capacity limits (maximum capacity for different Kubernetes resources within a single cluster).

    • ACK cluster quota limits and quota limits for ACK-dependent cloud products (such as ECS and VPC), and how to increase quotas.

FAQ

Can I create non-Auto Mode node pools in an Auto Mode cluster?

Yes, you can create self-managed node pools as needed. You can flexibly choose hosting configurations during creation. For more information, see Create and Manage Node Pools. For differences in hosting capabilities, see Comparison of Hosting Capabilities.