Overview

To automate software deployment with a Compute Nest service, service providers must follow these steps when creating the service:

• Create the appropriate artifact for your software deployment scenario. (Required)

• Configure a ROS deployment template based on your resource requirements and associate the artifact with it. (Required)

• Configure parameter plans. (Optional)

• Configure parameter mappings. (Optional)

Create an artifact

An artifact is a general term for a software deployment medium, including ECS images, Helm charts, software packages, and container images.

First, determine the deployment method for your software, and then create the corresponding artifact. For example, if your software is deployed on ECS instances, we recommend using an ECS image artifact. If your deployment relies on scripts, a file artifact is a suitable choice. For more information, see Manage artifacts.

Configure a ROS deployment template

Compute Nest templates are ROS templates. You can write a stack template that follows the ROS template specification. In the template, you define the required cloud resources, such as ECS instances and ApsaraDB RDS instances, and the dependencies between them. The ROS orchestration engine automatically creates and configures all resources based on the template, enabling automated deployment and operations. Resource Orchestration Service is a service provided by Alibaba Cloud to simplify cloud resource management. For more information, see What is Resource Orchestration Service?.

Associate an artifact

After you create an artifact, you must associate it with the ROS template to automate the deployment of both resources and software. In the service's ROS template, define a placeholder for the artifact. In the Artifact Association section, you are prompted to associate a specific artifact. When you deploy a service instance, Compute Nest identifies the placeholder in the template and automatically replaces it with the corresponding version of the artifact. For more information, see Associate an artifact.

Parameter plans

If a service has many parameters, you can create parameter plans to bundle parameters into selectable presets. This reduces the learning curve and helps prevent configuration errors. You can configure multiple parameter plans to provide users with best-practice configurations for different scenarios. For more information, see Parameter plan settings.

Automatic parameter mapping

For a fully managed service with dependent template parameters, you can use parameter mapping to set a parameter to a fixed value based on specific conditions. When you deploy a service instance, Compute Nest automatically applies the parameter values that you configured if the conditions are met. For more information, see Configure parameter mapping.

Other deployment capabilities

Compute Nest provides service providers with a dedicated deployment page for their customers. This page integrates features for displaying available parameters, providing price inquiries, and checking dependencies and service activation status:

• Available parameters: To address differences in stock and supported instance specifications across availability zones for Alibaba Cloud products, the dedicated deployment page dynamically links to Alibaba Cloud's resource inventory. It shows only the availability zones and instance specifications that have available stock, ensuring that any configuration a customer selects can be successfully provisioned. For example, a TiDB cluster deployment page provides a specification comparison for four plans: Basic edition - Standard subscription, Basic edition - Advanced subscription, High-spec edition - Standard subscription, and High-spec edition - Advanced subscription. It displays the instance specifications for components such as TiFlash, TiKV, PD, and TiDB. In the Topology section, users can set available parameters such as the number of TiDB Servers and PD Servers. The Basic Resource Configuration section includes required fields like VSwitch Availability Zone (you can select from availability zones G, H, I, and L; we recommend choosing a zone different from those used by other VSwitches), Virtual Private Cloud (VPC) Instance ID, and VSwitch Instance ID. Users can select from existing resources or click the provided link to create a new VPC and VSwitch.

• Price inquiry: Compute Nest provides a unified price inquiry interface. For private deployment scenarios, it displays software fees and costs for dependent resources. For managed service scenarios, it calculates the total service cost. The deployment page shows a Price Preview that includes software fees (e.g., ¥15,000.00) and Alibaba Cloud resource costs (e.g., ¥17.517/hour for pay-as-you-go). A resource details table lists each resource's type, purpose, specification, region, quantity, billing method, and estimated cost. For example: one Server Load Balancer (SLB) instance (slb.s3.small) at ¥1.27/hour; ECS instances (1 ControlServer at ecs.c6.large, 3 TiKVServers at ecs.i3.2xlarge, 2 TiDBServers at ecs.c6.xlarge, and 3 PDServers at ecs.c6.large); and an Elastic IP (EIP) at ¥0.02/hour plus traffic fees of ¥0.8/GB. All listed resources are billed on a pay-as-you-go basis.

• Dependency and service activation check: In the Dependency Check panel, all 5 dependencies are checked and reported as normal. Dependencies include snapshot, Auto Scaling (ESS) (which requires activating the AliyunServiceRoleForAutoScaling service-linked role), and Prometheus (which requires activating the AliyunServiceRoleForARMS service-linked role). All required role permissions are confirmed as activated.