Skip to content

My Deployments ​

Feature Overview ​

ItemContent
Applicable RolesModel Provider
Navigation PathModel Services > Studio > My Deployments
Page Route/modelone/my-deployments/models
Managed ObjectsOn-Cloud deployments, On-Prem deployments, deployment status, and publishing entry

Beginner Explanation ​

My Deployments works like a model deployment register. On-Cloud shows cloud deployments and On-Prem shows deployments in self-managed environments. Confirm the deployment type before reviewing status, resources, and call information.

Terminology ​

TermDescription
On-CloudA deployment that runs on cloud-platform resources.
On-PremA deployment that runs in self-managed physical or virtual infrastructure.
Deployment StatusThe current creation, running, failed, or stopped state of a deployment.
Call InformationThe URL, protocol, and authentication requirements used after deployment.

Open the relevant On-Cloud or On-Prem tab, locate the deployment, and review its details. To publish a new model, use the publishing entry and confirm the region before continuing.

Beginner Checklist ​

ScenarioDo FirstDo Not Do Directly
Unsure where a deployment runsCheck both deployment tabsQuery one tab repeatedly
Deployment status is abnormalReview details and status informationRedeploy immediately
Preparing to record call informationUse URL and credential placeholdersPut real credentials in documentation
Preparing to publish a modelConfirm region and resources firstSubmit publication immediately

Prerequisites ​

  1. The current account has permission to view Studio > My Deployments.
  2. At least one deployment record is listed on the page.
  3. The target deployment meets the conditions for displaying the publish entry, and Publish is visible in the more actions menu.
  4. Before publishing, the publish region, visibility scope, billing configuration, and call configuration risks have been confirmed.
  5. Before publication, confirm the deployment region, resources, model source, visibility, billing, and rate limits.

High-Risk Operation Boundary

Publishing a model affects model visibility, call methods, billing, and user access. Before submission, verify the publishing region, business scope, resources, model source, billing, and rate limits.

Page Description ​

The page separates deployment locations into On-Cloud and On-Prem. Lists show models, regions, status, and actions. The publishing entry starts a new model publication flow.

Page screenshots:

My Deployments list

The "My Deployments" page defaults to the On-Cloud model deployments list, displaying deployment name, associated model, cloud platform, running status, and management actions.

Main Operations ​

View On-Cloud Deployments ​

  1. Go to Model Services > Studio > My Deployments, which defaults to the On-Cloud tab.
  2. Filter cloud deployment instances by deployment name, model, or status.
  3. Verify deployment name, model type, cloud platform provider, deployment region, and current running status.
  4. In the action column, start, stop, view details, or initiate model publishing.

View On-Prem Deployments ​

  1. Click the "On-Prem" tab on the My Deployments page.
  2. The list displays model instances deployed on local private resource pools.
  3. Verify compute nodes, GPU card counts, container health status, and service endpoint addresses for target deployments.
  4. Manage the lifecycle or publish ready instances using the action column.

View On-Prem deployments

Switch to the On-Prem tab to inspect local model deployments on private resource pools, including GPU quota and service readiness status.

Publish a Model ​

  1. Click "Publish" in the action menu of a target running deployment card.
  2. In the dialog, select the publishing region (Private or Public Region) to determine visibility.
  3. Complete the service details, API key, billing mode, and rate limits, then submit.
  4. Return to My Models > My Published to verify the published model service status.

Publish Model entry

Click "Publish" in the deployment action menu to start the model publication workflow.

Choose where to publish

Select whether to publish the model to the tenant's Private Region or the platform's Public Region.

Publish Model page

Verify the auto-filled service endpoint URL, enter access credentials, configure billing and rate limits, and submit.

Parameter Reference ​

Field NameRequiredField TypeExampleDescription
Deployment NameYesTextExample DeploymentIdentifies the target deployment record.
Model NameYesTextExample ModelModel associated with the target deployment.
Deployment StatusYesStatus tagRunningIndicates whether the deployment may show a publish entry or meet later publishing conditions.
RegionYesTextChina (Shanghai)Region where the target deployment is running. Verify it against the publish region and business scope before publishing.
Resource SpecificationYesTextNVIDIA A10 x 1Shows GPU, CPU, memory, and other resource specifications used by the deployment.
Publish EntryYesMore actionsPublishEntry from the target deployment to the publish region selection flow.
Publish RegionYesCard selectionPrivate / PublicDetermines the publish target and visibility scope after the redirect.
Redirect TargetYesPage redirectMy Models > Publish ModelTarget page after a publish region is selected.
Publish ScopeConditionally requiredPage configurationPrivate / PublicConfirms the model visibility scope on the publish model page.
Billing ConfigurationConditionally requiredStep configurationToken / FreeConfirms pricing, free quota, or billing method on the publish model page.
Call ConfigurationConditionally requiredStep configuration<BASE_URL><ENDPOINT_PATH>Confirms request URL, API Key, model source ID, rate limits, and other call-related settings.
ActionsNoRow buttons / more menuStart / Stop / Details / PublishPage entries for viewing, controlling, or entering the publish flow.

Pitfalls ​

  • My Deployments shows deployment status. It does not mean the model has been published to the marketplace.
  • When deployment is abnormal, check Resource Pool, Runtime Image, model asset, and startup logs before submitting again.
  • Deleting, stopping, or restarting a deployment may affect Model Consumers. Confirm traffic and rollback path first.

Result Validation ​

Check ItemSuccess SignalIf Abnormal
Page is accessibleThe My Deployments page opens, and the On-Prem and On-Cloud tabs are visible.Check account permissions, navigation path, and page loading status.
Deployment list loadsThe target deployment card shows deployment name, model name, status, region, and resource specification.Click "Search" or Reset and retry. Check filters and deployment permissions if needed.
Target deployment status is visibleThe target deployment status is shown as the real page status, such as Running.Confirm whether the deployment task is complete, or open Details to view deployment status.
Publish entry is visibleEligible deployments show Publish in the more actions menu.Check deployment status, account permissions, and row actions.
Publish region can be selectedThe Publish Model dialog opens and shows Private, Public, and the corresponding publish buttons.Close the dialog and retry, or check whether the account has permission for the selected publish region.
Redirect target is correctAfter selecting a publish region, the page opens the Publish Model page under Model Services > Studio > My Models.Check publish region permissions, page route, and browser redirect status.
Publish fields are visibleThe page shows Basic Information, Billing Configuration, Rate Limit Configuration, and their key fields.Go back and select the publish region again, or refresh the publish model page.

FAQ ​

Deployment List Is Empty ​

Symptom:

The target deployment is missing from On-Cloud or On-Prem.

Possible Causes:

  • The wrong tab is open.
  • Filters or visibility do not match.

Resolution:

  1. Check both deployment tabs.
  2. Reset filters and verify permission.

Deployment Details Do Not Open ​

Symptom:

Opening a deployment returns no details or an error.

Possible Causes:

  • The record is still being created.
  • The record expired or permission is missing.

Resolution:

  1. Refresh the list and verify status.
  2. Confirm the record exists and the account can view it.

Deployment Status Does Not Change ​

Symptom:

Status remains in creating or processing.

Possible Causes:

  • Resources are not ready.
  • The deployment task encountered an error.

Resolution:

  1. Review stages and errors in details.
  2. Ask an authorized operator to inspect the task.

Call Information Is Unavailable ​

Symptom:

No call information is listed after deployment.

Possible Causes:

  • The service is not ready.
  • Call permission or network scope is not configured.

Resolution:

  1. Verify deployment health.
  2. Check authorization and network access.

Model Publication Fails ​

Symptom:

No deployment record is created after submission.

Possible Causes:

  • Resource, source, or billing configuration is incomplete.
  • The name is duplicated or the region is incorrect.

Resolution:

  1. Correct fields indicated by the page.
  2. Verify name and region before resubmission.

Notes ​

  • Publishing a model affects model visibility, call methods, billing configuration, and user access.
  • Selecting the wrong publish region may publish the model to the wrong site, region, or business scope.
  • 发布 / Publish, 提交 / Submit, and 保存 / Save are high-risk final actions.
  • Do not write real accounts, secrets, tokens, AK/SK, Endpoints, API Keys, customer names, pricing strategies, cloud resource IDs, or internal test parameters.

Next Steps ​

  1. Return to My Models to view the publish model configuration progress.
  2. Verify the model visibility, billing configuration, and call configuration for the selected publish region.
  3. If a real publish action has been performed, open model details or the call page to confirm status and access control.