SYSTEM_CONSOLE v2.4.0

Data Mesh

LAST_UPDATED: 2025-09

Data Mesh shifts data ownership to the business domains that produce and consume it. The model only works when each domain treats its datasets as products: with documented schemas, SLOs, and a named owner on call. Without that, you get distribution without accountability.

Key Takeaways

  • 01 Data is a product, not a byproduct of application code.
  • 02 Domain teams own the full lifecycle of their data products.
  • 03 Platform teams provide the 'paved road' for autonomy.
  • 04 Federated governance makes data products interoperable across domains.

Checklist

  • Data product specification filled for each new product.
  • Domain owner and on-call rotation clearly defined.
  • Standard governance tags applied to all datasets.
  • Self-service provisioning templates available.

What is a data product?

A data product is a high-quality, ready-to-use dataset that is discoverable, addressable, trustworthy, and self-describing. It must include:

Technical Core

  • Schema (standardized format)
  • Access policy (IAM roles)
  • SLOs (Freshness, Completeness)

Product Wrapper

  • Documentation (What does this mean?)
  • Support model (Who do I contact?)
  • Change & Deprecation policy

Ownership model

Data mesh ownership model

Onboarding playbook

1. Template Provisioning

Domain team uses the platform's CI/CD templates to create a new ingestion and transformation project.

2. Contract Review

The proposed data schema is reviewed against federated standards for naming and common data types.

3. Observability Setup

Baseline monitors for pipeline heartbeats and freshness are automatically enabled via the platform templates.

Failure modes

  • ! "Mesh" becomes Chaos: Lack of strong central standards leads to every domain using different tools and formats.
  • ! No Standards: Domains produce data products that cannot be easily joined together (e.g., different date formats).
  • ! Platform Bottleneck: The platform team still manually reviews every change, defeating the purpose of decentralized ownership.
  • ! Unclear Ownership: A shared data product breaks and no single team feels responsible for fixing it.