Container Service for Kubernetes:Auto Mode overview

Auto Mode provides the following benefits:

• Comprehensive managed O&M, including a fully managed cluster control plane and key system add-ons, automatic Kubernetes version upgrades, and full node lifecycle management.

• Automatic node scale-in and scale-out, without manual capacity planning, to help optimize node resource costs.

• A built-in managed Application Load Balancer (ALB) Ingress that provides a unified application access point for Services in the cluster.

• Security by default, with an immutable node host file system (FS) that enhances system security.

• Full compatibility with the native Kubernetes ecosystem, including support for resources such as DaemonSet.

Auto Mode is ideal for general containerized scenarios, especially in the following cases:

• You have limited resources and cannot dedicate a large team to manage Kubernetes infrastructure.

• You lack professional Kubernetes O&M experience but want to use Kubernetes to containerize and modernize your applications.

• You need to quickly deploy standardized Kubernetes environments worldwide.

Features

• Managed O&M

  • Cluster level: The cluster control plane and key system add-ons, such as CoreDNS and Container Storage Interface (CSI), are fully managed. Automatic Kubernetes version upgrades are enabled by default.

  • Node level: By default, ACK creates a system node pool named ack-system (System node pool) to host cluster system add-ons, and a node pool named default (Intelligent managed node pools) to host your business workloads. ACK automatically recommends the optimal instance types, eliminating the need for manual configuration. ACK also handles O&M tasks such as OS and node add-on upgrades and node security vulnerability patching.

• Elastic node scaling

  • Auto Mode includes built-innode instant scaling, which dynamically scales nodes based on workload requirements. This eliminates upfront capacity planning and optimizes resource costs.

• Second-level pod scaling

  • Auto Mode supportssecond-level container scaling with ack-autoscaling-placeholder, which provides a buffer for pod autoscaling. This feature allows nodes to be pre-warmed, ensuring rapid pod scale-out without the delay of creating and joining new nodes.

• Application Load Balancer (ALB)

  • A built-in ALB Ingress, based onALB, provides enhanced Ingress traffic management, is compatible with Nginx Ingress, handles complex business routing, supports automatic certificate discovery, and supports the HTTP, HTTPS, and QUIC protocols.

• Security

  • Optimized base software stack by default: Nodes useContainerOS. Its immutable root file system enhances security, while its streamlined system and configuration accelerate node startup. The optimized kernel helps maximize hardware performance.

  • Security and observability by default: Features such as cluster check / cluster inspection,API server auditing, andcontrol plane add-on log collectionare enabled by default, helping you quickly identify and resolve cluster issues.

Usage notes

The following usage notes apply to Auto Mode clusters:

• Manual node O&M — When ACK performs O&M tasks such as OS upgrades, software upgrades, or security vulnerability patching, the tasks involve software version changes, configuration modifications, reboots, and node draining.

  Important

  Do not perform manual O&M on the ECS instances in the node pool, such as rebooting, mounting data disks, or logging in to nodes to modify configurations. Manual operations can conflict with automated policies.

• Workload availability — To ensure that nodes can be drained for maintenance without business interruption, configure your workloads with an appropriate number of replicas, PreStop hooks for graceful shutdown, and PodDisruptionBudget policies.

• Storage — Nodes may be automatically replaced by the elastic scaling policy.

  Note

  Avoid storage methods that rely on specific node paths, such as hostPath. Use a persistent volume claim (PVC) for persistent storage.

• ContainerOS compatibility — Because ContainerOS has an immutable root file system, if your applications require node environment customization or local data persistence, seeContainerOSto assess their compatibility in advance.

Get started

1. Create an ACK Auto Mode cluster.

2. Deploy a workload in the cluster and expose it to the internet by using an ALB Ingress. For more information, seeDeploy a workload and configure load balancing.

3. Create more node pools (with or without Auto Mode) as needed. For more information, seeCreate and manage a node pool.

Quotas and limitations

• When you use an intelligent managed node pool, ACK dynamically scales nodes based on workload demand. By default, it supports scaling out to a maximum of 50 nodes. You can change the maximum number of instances by using the node pool's scaling settings.

• Intelligent managed node pools do not support certain instance types, such as Arm-based instances or instances with local disks, and they support only ContainerOS 3.6 and later. ACK recommends default instance type families that meet the needs of most use cases. You can also adjust the settings in the console based on your specific business requirements. We recommend configuring a sufficient number of instance types to enhance the node pool's elasticity and prevent scaling failures.

• If you have a large-scale cluster or a large number of resources under your account, you must follow the quotas and limits that apply to ACK clusters.

  • ACK configuration limits, such as account balance requirements, and single-cluster capacity limits, which specify the maximum capacity for different Kubernetes resources in a single cluster.

  • Quotas for ACK clusters and dependent cloud services such as ECS and VPC, and information about how to increase a quota.

Billing

You can enable Auto Mode only forACK managed Pro cluster. The feature itself is free, but you are billed as usual for cluster management fees and associated cloud resources. For more information, seeBilling overviewandFees for cloud resources.

Shared responsibility model

Auto Mode is designed to provide automated and intelligent O&M for Kubernetes clusters, reducing your operational overhead. However, you are still responsible for certain tasks in some scenarios.

FAQ

labels:
  alibabacloud.com/system-nodepool: "true"
taints:
  - key: CriticalAddonsOnly
    value: "true"
    effect: NoSchedule