Scenario Overview - On Cloud Model Deployment and Calling

This scenario guides tenant users through quickly deploying a model from the cloud marketplace and obtaining API calling capability.

Applicable Roles

• Model Provider
• End User

Scenario Goals

• Add or confirm an available cloud account.
• Select a target model, inference engine, and compute plan to complete deployment.
• Obtain deployment state, monitoring, and API access from My Deployments.

Scenario Flow

Main path: Prepare a user access account -> Select a model and compute plan -> Check deployment state -> Call and manage the lifecycle

Stage	Key Result
1. Prepare the account	The user cloud account passes validation and authorized regions are visible
2. Create deployment	Model, engine, strategy, specification, and estimated cost are confirmed and submitted
3. Check runtime	The deployment reaches Running without repeated event or monitoring errors
4. Call and manage	The API request succeeds and cloud resources can be stopped or released as needed

Before You Start

• The Platform Operator has completed cloud-resource, business-region, and tenant authorization.
• The target model asset is published.
• A redacted example of the user cloud-account credential is prepared.

Recommended Reading Order

1. Add an access account.
2. Create a quick deployment.
3. Review state, events, monitoring, and API access in My Deployments.

Document Index

Document	Description
Access Accounts	Add and validate the user's own cloud account
Quick Deployment	Select an account and model and use the recommended deployment plan
My Deployments	Review state, details, API access, monitoring, events, and lifecycle actions

Related Scenarios

Platform Side - Upstream

• On Cloud Resource Access: confirms that the Platform Operator has connected the target cloud platform and completed business-region and tenant-cloud authorization
• On Cloud Model Asset Publishing: supplies the model assets selectable in Quick Deployment step 1

Platform Side - Downstream and Cross-Scenario

• Publish Models: after selecting Publish in My Deployments, complete basic information, billing, and rate limits to publish the deployment as a callable service
• My Deployments User Manual: reviews state, events, and monitoring for deployed instances

User Side - Downstream

• Model Experience and API Calling: explains how end users discover, try, and call the model after publication

Completion Checklist

Purpose: These are the scenario exit criteria. Use them to confirm that the result is observable and reviewable before continuing. If any check fails, return to the relevant feature guide and follow its troubleshooting section.

Check	Pass Criteria
1	The access account passes validation and authorized regions are visible.
2	Quick deployment can select the target model, inference engine, and compute plan.
3	Deployment reaches Running without repeated errors in events or monitoring.
4	A controlled request succeeds with redacted endpoint and API key handling.
5	Stopping or deleting a test deployment releases resources as expected.

Troubleshooting

Symptom	Check First
No account or region is available	Account validation, resource grants, and tenant scope
Recommendation has no plan	Model asset, framework, compute plans, and regional capacity
Deployment remains creating or fails	Events, account permission, image, quota, and cloud resource state
Deployment runs but calls fail	Endpoint, authentication, port, protocol, and service health