API standard and pre-built SDKs in multi-language
The OpenAPI specification of this product (pai-dlc/2020-12-03) follows the ROA standard. Alibaba Cloud provides pre-built SDKs for popular programming languages to abstract low-level complexities such as request signing. This enables developers to call APIs using language-specific syntax without dealing with HTTP details directly.
Custom signature
If your specific needs, such as a customized signature, are not supported by the SDK, manually sign requests using the signature mechanism. Note that manual signing requires significant effort (usually about 5 business days). For support, join our DingTalk group (ID: 147535001692).
Before you begin
An Alibaba Cloud account has full administrative privileges. A compromised AccessKey pair exposes all associated resources to unauthorized access, posing a significant security risk. To call APIs securely, create a Resource Access Management (RAM) user with API access only, configure its AccessKey pairs, and implement the principle of least privilege (PoLP) through RAM policies. Use the Alibaba Cloud account only when its permissions are explicitly required for specific scenarios.
Tasks
|
API |
Title |
Description |
| CreateJob | Create a job | Creates a job and runs it in a cluster. You can specify the datasource config, code source configuration, startup command, and compute resource configuration for each node on which the job runs. |
| UpdateJob | UpdateJob | Updates a job's configuration, such as its priority. |
| ListJobs | Get job list | Retrieves a list of jobs with support for pagination, sorting, and conditional filtering. |
| GetJob | Get job details | Retrieves the detailed configuration and runtime information of a job. |
Task template
|
API |
Title |
Description |
| CreateJobTemplate | CreateJobTemplate | Creates a job template. |
| GetJobTemplate | GetJobTemplate | Retrieves the details of a job template. |
| DeleteJobTemplate | DeleteJobTemplate | Deletes a job template. You cannot delete a job template that is in use by a job. |
| UpdateJobTemplate | UpdateJobTemplate | Updates a job template. |
| ListJobTemplates | ListJobTemplates | Lists job templates in a specified workspace. You can paginate, sort, and filter the results by creator, `TemplateId`, or `TemplateName`. |
| SetJobTemplateDefaultVersion | SetJobTemplateDefaultVersion | Sets the default version of a job template. |
TensorBoard
|
API |
Title |
Description |
| UpdateTensorboard | UpdateTensorboard | Update a Tensorboard. |
RayHistoryServer
|
API |
Title |
Description |
| CreateRayHistoryServer | CreateRayHistoryServer | Creates a RayHistoryServer. |
| GetRayHistoryServer | GetRayHistoryServer | Retrieves the details of a specific RayHistoryServer. |
| ListRayHistoryServers | ListRayHistoryServers | Listing RayHistoryServer resources |
| StartRayHistoryServer | StartRayHistoryServer | Starts a Ray History Server. |
| StopRayHistoryServer | StopRayHistoryServer | Stops a Ray History Server. |
| DeleteRayHistoryServer | DeleteRayHistoryServer | Deletes a Ray History Server. |
Other
|
API |
Title |
Description |
| GetDashboard | GetDashboard | Retrieves the Dashboard URL for a DLC job, if available. |
| GetRayDashboard | GetRayDashboard | Retrieve the Ray Dashboard URL. |
Others
|
API |
Title |
Description |
| CreateTensorboard | CreateTensorboard | Creates a TensorBoard by using a job or specifying a data source configuration. |
| DeleteJob | DeleteJob | Deletes a completed or stopped job. |
| DeleteTensorboard | DeleteTensorboard | Deletes a stopped TensorBoard. |
| GetJobEvents | GetJobEvents | Obtains the system events of a job. |
| GetJobMetrics | GetJobMetrics | Obtains the monitoring data of a job, including the CPU, GPU, and memory utilization, network, and disk read/write rate. ⚠️ Note: Except for pay-as-you-go tasks based on general-purpose computing resources, all task types are connected to CloudMonitor. Use the CloudMonitor API to call related monitoring. The overwritten features in the original API are no longer maintained. For more information, see \\[Training monitoring and alerting]\\(https://www.alibabacloud.com/help/zh/pai/user-guide/training-monitoring-and-alerting). |
| GetJobSanityCheckResult | GetJobSanityCheckResult | Obtains specified job sanity check result in a Deep Learning Containers (DLC) job. |
| GetMetrics | GetMetrics | Retrieve metrics data. |
| GetPodEvents | GetPodEvents | Obtains the system events of a specific node in a job to locate and troubleshoot issues. |
| GetPodLogs | GetPodLogs | Obtains or downloads the logs of a node for a task. The logs are from the stdout and stderr of the system and user scripts. |
| GetTensorboard | Get tensorboard details | Retrieves the details of a Tensorboard instance. |
| GetTensorboardSharedUrl | GetTensorboardSharedUrl | Obtains the shareable link of a TensorBoard task. The link contains digital tokens. You can use a shareable link to access a TensorBoard task. |
| GetToken | GetToken | Obtains the sharing token of a DLC job. This token is used to view the information about the shared job. |
| GetWebTerminal | GetWebTerminal | Provides methods and steps to obtain a HTTP link for accessing a container. |
| ListEcsSpecs | ListEcsSpecs | Queries the list of supported instance types. |
| ListJobSanityCheckResults | ListJobSanityCheckResults | Obtains the results of all sanity checks for a DLC job. |
| ListTensorboards | Query tensorboard list | Queries a list of created Tensorboard instances. |
| StartTensorboard | StartTensorboard | Starts a TensorBoard instance. |
| StopJob | StopJob | Stops a running job. |
| StopTensorboard | StopTensorboard | Stops a TensorBoard instance. |