Data Fabric

Responsibilities

• Define the data model boundaries and consistency strategy.
• Provide storage patterns (OLTP/OLAP/cache/search) with clear ownership.
• Control retention, privacy, and lifecycle policies.

Key interfaces

• Data access contracts (repositories, query APIs, schema versioning).
• Caching and invalidation mechanisms.
• Migration and backfill workflows with rollback strategy.

Operational signals

These are the measurements that tell you whether this layer is healthy in production.

• Read/write latency, lock contention, and connection pool saturation.
• Replication lag, queue/backlog size (if streaming).
• Cache hit rate and eviction churn.

Failure modes

• Hot partitions / uneven load distribution.
• Schema drift and broken migrations.
• Data leaks via over-broad access or missing redaction.

Production readiness checklist

• Define SLOs per datastore and enforce backpressure.
• Test migrations in shadow environments; keep rollbacks ready.
• Implement data classification and least-privilege access controls.