Skip to main content

Policies

Create and manage governance policies to enforce compliance, security, and quality standards across your platform.

How to Create a Policy​

Step 1: Navigate to Governance​

  • Click Governance in the main navigation
  • Click Create Policy button

Step 2: Define Policy Basics​

Policy Metadata:

  • Name: Descriptive policy name (e.g., "Production Deployment Approval")
  • Description: Purpose and scope of policy
  • Version: Semantic version (e.g., 1.0.0)
  • Category: Security, Compliance, Quality, or Operational
  • Severity: Critical, High, Medium, Low, or Info
  • Applicable Resources: Solution, Workflow, Application, Model, Dataset, API

Step 3: Configure Stages​

Policies consist of sequential or parallel stages that resources must complete. Each stage can have:

Stage Fields​

Define custom form fields for each stage:

  • Field Types: text, textarea, richtext, radio, checkbox, dropdown, multiselect, file upload, number, date, datetime, URL, email
  • Validation: Required fields, regex patterns, min/max values, custom validators
  • Conditional Display: Show/hide fields based on other field values
  • Help Text: Guidance for completing each field

Stage Validators​

Assign approvers to each stage:

  • Validator Types: Specific user, role (e.g., Admin), or group
  • Approval Logic (for groups): All must approve, any one can approve, or majority vote
  • Delegation: Allow validators to delegate to others
  • Escalation Rules: Auto-escalate if not approved within X hours
  • Notifications: Email, in-app, or digest notifications

Stage Configuration​

  • Order: Sequential position in workflow
  • Dependencies: Which stages must complete first
  • SLA: Expected completion time in hours
  • Required: Can this stage be skipped?
  • Parallel Execution: Run simultaneously with other stages

Step 4: Set Enforcement Rules​

Define how policy violations are handled per environment:

ActionBehavior
Hard BlockResource cannot be deployed/promoted until policy completes
Soft BlockWarn user, but allow override with justification
WarningLog warning, allow deployment to proceed
Post ReviewAllow deployment, require policy completion afterward

Step 5: Activate Policy​

  1. Review all stages and validation rules
  2. Save as Draft to test with sample resources
  3. When ready, toggle Active to enforce policy
  4. Share with specific users or teams who should see this policy

Policy Examples​

Production Deployment Approval​

Purpose: Require approval before deploying to production

Stages:

  1. Security Review: Security team validates security requirements
  2. Technical Approval: Lead engineer approves technical implementation
  3. Business Approval: Product owner confirms business requirements met

Enforcement: Hard Block in Production, Warning in Development

Data Privacy Compliance​

Purpose: Ensure data handling complies with privacy regulations

Stages:

  1. Data Classification: Classify data sensitivity level
  2. Privacy Impact Assessment: Document privacy implications
  3. Legal Review: Legal team approves data handling procedures

Enforcement: Hard Block in all environments

Model Validation​

Purpose: Validate ML model performance before deployment

Stages:

  1. Performance Metrics: Document model accuracy, precision, recall
  2. Bias Analysis: Assess model for bias across protected groups
  3. Model Approval: Data science lead approves deployment

Enforcement: Hard Block in Production, Soft Block in Staging

Important

Policies with "Hard Block" enforcement will prevent resource deployment until all stages are completed. Use carefully in development environments.