All Products
Search
Document Center

Container Service for Kubernetes:Create an Argo workflow cluster

Last Updated:Jun 17, 2026

Argo workflow clusters run on Alibaba Cloud Container Compute Service (ACS) in serverless mode with BestEffort QoS, optimized for elastic, cost-efficient scheduling of large-scale workflows.

Prerequisites

Create a workflow cluster

  1. Log on to the Argo Workflow Clusters console.

  2. Click Create Workflow Cluster. Configure the parameters and click Next. After dependency checks pass, click Create Cluster.

    Parameter

    Description

    Cluster Name

    The cluster name.

    1–63 characters: letters, digits, underscores (_), or hyphens (-). Must start with a letter.

    Region

    The region to deploy the cluster.

    VPC

    An existing VPC.

    vSwitch

    An existing vSwitch.

    Resource Group

    The resource group for permission management and cost allocation.

    A resource can belong to only one resource group.

    Tags

    Key-value pair tags to identify the cluster.

    API Server Access

    No configuration required. A pay-as-you-go private network Classic Load Balancer (CLB) instance is automatically created as the internal endpoint for the API Server.

    Important

    Do not delete this instance, or you will lose access to the cluster API Server.

    Create and Attach EIP

    If enabled, an elastic IP address (EIP) is attached to the CLB instance, enabling public network access to the API Server.

    Enable Components and Audit Log

    Enables API Server audit logging to collect Kubernetes API requests and results. To enable later, see Collect container logs from ACK clusters and Use the API Server audit feature.

Delete a workflow cluster

Important

Before deleting an Argo workflow cluster, delete all workflows in the cluster to release ACS resources.

  1. Log on to the Argo Workflow Clusters console.

  2. In the Actions column, click image > Delete Cluster. In the dialog box, click OK.

References

To use the cluster, enable Argo Server and access the workflow console.