Transformation & Processing → Event Schema & Registry

Governance Rules

The policy layer on top of the schema registry — who can register what, how PII fields must be tagged, and what naming conventions apply.

High-Level Design

Governance Rules make downstream privacy controls actually enforceable.

Data Source
Compatibility Checking
Runs alongside this policy check
→
Ingestion
Schema Registration Request
Subject to governance review
→
Processing
Governance Rules
Schema policy enforcement
→
Foundation
Classification Taxonomy
Governance & Security's source of valid PII tags
→
Intelligence
Audit Logs
Every rejection is logged
→
Activation
Enforceable Downstream Privacy Controls
Consent enforcement depends on correct tagging here

💼 Business Context

  • Prevents schema sprawl (ten slightly different ways to represent 'order total') that makes cross-team analytics painful
  • Ensures every PII field is correctly classified at the source, which is what makes downstream privacy controls (consent enforcement, access control) actually reliable
  • Owned jointly by Platform Engineering and Data Governance

🔌 Technical Overview

Governance Rules are policy checks the Schema Registry runs on every registration or change: naming-convention enforcement (snake_case, consistent event-name verbs), mandatory classification tags on any field that could contain PII, and required ownership metadata (which team owns this event type). Registrations that violate policy are rejected with a specific, actionable error rather than silently accepted.

Enforced Rules

Naming conventions Mandatory PII classification tags Required team ownership metadata Deprecation notice period

💾 Governance Policy Check

{
  "field": "properties.customer_email",
  "issue": "missing PII classification tag",
  "required_action": "add classification: 'pii.email'",
  "status": "rejected"
}

🔗 Integration Points

  • Schema Registry — where governance checks run alongside compatibility checks
  • Governance & Security's data classification taxonomy — the source of valid PII tags
  • Data Governance team — reviews and updates the rule set over time
  • Audit Logs — every governance rejection is logged for compliance visibility

🧰 Services Consumed

  • Owning microservice — Cxos.Processing.SchemaGovernance.Api (see the Full Application Service Map)
  • Database — Azure Database for PostgreSQL (relational, ACID)

⚠️ Non-Functional Considerations

  • Scale: policy checks are rule-based and fast, no different in cost from compatibility checking
  • Latency: runs synchronously as part of schema registration, same CI-gate pattern as compatibility checking
  • Reliability: rule updates are versioned so historical schemas aren't retroactively marked non-compliant
  • Security/Privacy: this is itself a core privacy control — correct PII tagging here is what makes consent enforcement and encryption policy enforceable downstream

🎯 Enterprise Example

A team registers a new event type with a customer_ssn field but forgets to tag it as PII. Governance Rules rejects the registration until the field is properly classified — preventing an unclassified sensitive field from ever reaching the pipeline where it could bypass consent and encryption controls built around classification tags.

← Back to Event Schema & Registry