Control-M for Azure Machine Learning

Azure Machine Learning enables you to build, train, deploy, and manage machine learning models on premises, in the cloud, and on edge devices.

Control-M for Azure Machine Learning enables you to do the following:

Execute an Azure Machine Learning endpoint pipeline and stop, start, restart, or delete a host or cluster.
Manage Azure Machine Learning credentials in a secure connection profile.
Connect to any Azure Machine Learning endpoint.
Introduce all Control-M capabilities to Control-M for Azure Machine Learning, including advanced scheduling criteria, complex dependencies, resource pools, lock resources, and variables.
Integrate Azure Machine Learning jobs with other Control-M jobs into a single scheduling environment.
Monitor the status, results, and output of Azure Machine Learning jobs.
Attach an SLA job to the Azure Machine Learning jobs.
Run 50 Azure Machine Learning jobs simultaneously per Agent.

Control-M for Azure Machine Learning Compatibility

The following table lists the prerequisites that are required to use the Azure Machine Learning plug-in, each with its minimum required version.

Component	Version
Control-M/EM	9.0.21.300
Control-M/Agent	9.0.21.300
Control-M Application Integrator	9.0.21.300
Control-M Automation API	9.0.21.301

Control-M for Azure Machine Learning is supported on Control-M Web and Control-M Automation API, but not on the Control-M client. You cannot use Promotion with the plug-in.

To download the installation files for each prerequisite, see Obtaining Control-M Installation Files.

Setting up Control-M for Azure Machine Learning

This procedure describes how to deploy the Azure Machine Learning plug-in, create a connection profile, and define an Azure Machine Learning job in Control-M Web and Automation API.

Integration plug-ins released by BMC require an Application Integrator installation. However, these plug-ins are not editable and you cannot import them into Application Integrator. To deploy these integrations to your Control-M environment, import them directly into Control-M with Control-M Automation API.

Before You Begin

Verify that Java is installed, as described in Control-M External Java Installation.
Verify that Automation API is installed, as described in Automation API Installation.

Begin

Create a temporary directory to save the downloaded files.
Download the Azure Machine Learning plug-in from the Control-M for Azure Machine Learning download page in the EPD site.
Do one of the following to install the Azure Machine Learning plug-in:
- (9.0.21 or higher) Do the following to run the Control-M Automation API provision service:
  1. Log in as an administrator to the Control-M/EM server host and save the downloaded ZIP file to the following location:
    - Linux: $HOME/ctm_em/AUTO_DEPLOY
    - Windows: <EM_HOME>\AUTO_DEPLOY
  2. Log in to the account where Control-M/Agent is installed and run the following provision image command:
    - Linux: ctm provision image ZML_plugin.Linux
    - Windows: ctm provision image ZML_plugin.Windows
- (9.0.20.200 or lower) Run the Control-M Automation API deploy service, as described in deploy jobtype.
Create an Azure Machine Learning connection profile in Control-M Web or Automation API, as follows:
- Web: Create a Centralized Connection Profile with Azure Machine Learning Connection Profile Parameters
- Automation API: ConnectionProfile:Azure Machine Learning
Define an Azure Machine Learning job in Control-M Web or Automation API, as follows:
- Web: Create a Job with Azure Machine Learning Job parameters
- Automation API: Job:Azure Machine Learning

To remove this plug-in from an Agent, see Removing a Plug-in from the Agent. The plug-in ID is ZML022023.

Change Log

The following table provides details about changes that were introduced in new versions of this plug-in:

Plug-in Version	Details
1.0.02	Added Executes a Job action to job definition. Added Job ID to job definition.
1.0.01	Internal changes to post-API calls
1.0.00	Initial version