Design and Govern Event Schemas

Event Schema Registry Expert designs and governs schemas for event-driven architectures: Avro and JSON Schema design with domain-based naming and semantic versioning, schema-registry integration through a Confluent client publishing Avro-serialized events, compatibility-type selection across BACKWARD, FORWARD, FULL, and NONE with worked backward-compatible evolution examples, registry deployment via Docker Compose, direct Schema Registry API usage, and per-environment compatibility strategy plus performance tuning.

Use it when designing or evolving event schemas for a Kafka-based or similar event-driven system - writing an Avro or JSON Schema for a new event type, choosing a compatibility level for a subject, adding a field to an existing schema without breaking consumers, standing up a Confluent Schema Registry instance, or calling its REST API directly to register or check a schema. It's specific to schema-registry-based event governance - it's not a general API-design or data-modeling agent, and it assumes Kafka or an equivalent event backbone is already in place.

Input is the event type to model and the compatibility guarantees needed. Output is schema and integration code: an Avro record schema (UserRegistered) with a domain namespace, field docs, and a nullable metadata map; a JSON Schema (draft 2020-12) for an OrderCreated REST event with required fields, nested $defs, and format constraints; a Python EventPublisher class that fetches the latest schema from a Confluent Schema Registry client, builds an AvroSerializer, and publishes a Kafka event; a side-by-side Product schema evolution example adding an optional category field in a backward-compatible way; a docker-compose.yml standing up cp-schema-registry with BACKWARD compatibility against a Kafka backend; curl examples registering a schema, fetching its latest version, and checking compatibility against the registry's REST API; and a SchemaRegistryConfig class selecting NONE, BACKWARD, or FULL compatibility per environment for development, staging, and production.

Built around Confluent Schema Registry and Kafka, with Avro and JSON Schema formats, a Python confluent_kafka client, and Docker Compose for local registry deployment:

// Version 1.1.0 - Backward compatible addition
{
  "type": "record",
  "name": "Product",
  "fields": [
    {"name": "id", "type": "string"},
    {"name": "name", "type": "string"},
    {"name": "price", "type": "double"},
    {"name": "category", "type": ["null", "string"], "default": null}
  ]
}

Platform and backend engineers governing event schemas across services who want the full lifecycle covered - naming conventions, a schema-review and CI/CD compatibility-check process, deprecation policies for old versions, usage monitoring, local schema caching to cut registry calls, and circuit breakers for registry unavailability - not just a one-off schema definition.