Setting up Watson OpenScale

To set up IBM Watson OpenScale, you can connect to databases, create access policies, and manage users and roles.

Before you begin

After provisioning Watson OpenScale, you can complete the following setup:

  1. Connecting an existing or new database
  2. Creating access policies
  3. Managing users and roles

Connecting an existing or new database

Watson OpenScale offers a free database for Lite plans or you can connect Databases for PostgreSQL or Db2 databases. If you are using a standard v2 plan, it is recommended to use your existing database or purchase a new database.

Complete the following steps to connect an existing or new database:

  1. From the Watson OpenScale navigation menu, click Configure Watson OpenScale configure icon.

  2. Click Database and select a database type. If an existing database cannot be located, you can create a new database. Select your database screen displays with fields for you to enter database type and database location

    The free Lite plan database has some important limitations:

    • The free database is hosted, and is not directly accessible to you.
    • Watson OpenScale has full access to your database and data.
    • If you reconfigure the database, Watson OpenScale loses your data.
    • The free database is not GDPR-compliant. If your model processes personally identifiable information, you must purchase a new database or use an existing database that is GDPR compliant.
  3. Select a database location. To connect a database outside of your IBM Cloud account, click Other and provide connection information.

  4. Save your database connection and all changes made.

Supported Db2 databases for data mart

  • Db2 on Cloud with default settings
  • Db2 Warehouse on Cloud with default settings
  • Db2 with the following settings:
    • Product name: " Db2 Enterprise Server Edition"
    • Product identifier: "db2ese"
    • Version information: "11.5 or later"

Creating access policies

  1. From the IBM Cloud homepage, click Manage > Access (IAM).
  2. From the IAM dashboard, click Users and select Invite user.
  3. Complete the following fields:
    • How do you want to assign access? : Access policy.
    • Which service do you want to assign access to? : watsonx.governance and All IAM account management service then click Next.
    • How do you want to scope the access : Assign Viewer access to All IAM account management service and for the watsonx.governance policy, then click Next.
      • If you select Specific resources, select an attribute type and specify a value for each condition that you add.
      • If you select Service instance in the Attribute type list, specify your instance in the Value field.
  4. If you have multiple instances, you must find the data mart ID to specify the instance that you want to assign users access to. You can use one of the following methods to find the data mart ID:
    • On the Insights dashboard, click a model deployment tile and go to Actions > View model information to find the data mart ID.
    • On the Insights dashboard, click the navigation menu on a model deployment tile and select Configure monitors. Then, go to the Endpoints tab and find the data mart ID in the Integration details section of the Model information tab.
  5. Select the Reader role in the Service access list.
  6. Assign access to users.
    • If you are assigning access to new users, click Add, and then click Invite in the Access summary pane.
    • If you are assigning access to existing users, click Add, and then click Assign in the Access summary pane.
Note:

You can create an access group with the required permissions for watsonx.governance and assign users to the group. For details on creating an access group, see Managing users and access.

Watson OpenScale users and roles

You can assign roles to Watson OpenScale users to collaborate on model evaluations. For more details, see Managing users and access.

The following table lists permissions for roles that you can assign for access to evaluations. The Operator and Viewer roles are equivalent.

Table 1. Operations by role
The first row of the table describes separate roles that you can choose from when creating a user. Each column provides a checkmark in the role category for the capability associated with that role.
Operations Admin role Editor role Viewer/Operator role
Add machine learning engine configuration
Remove machine learning engine configuration
Update machine learning configuration
View machine learning configuration
Add database configuration
Remove database configuration
Update database configuration
View database configuration
Model approval
Evaluation
View users and roles
Add subscription to dashboard
Remove subscription from dashboard
View subscription
Configure monitoring condition
View monitoring condition
Upload payload logging record
Upload feedback data
Upload training data CSV file in model risk management
Run auto setup
API calls to update the system
API calls to query the subscriptions and monitoring

Parent topic: Setting up the platform for administrators