confluent-schema-registry

# Confluent Schema Registry Skill

Expert knowledge of Confluent Schema Registry for managing Avro, Protobuf, and JSON Schema schemas in Kafka ecosystems.

## What I Know

### Schema Formats

**Avro** (Most Popular):
- Binary serialization format
- Schema evolution support
- Smaller message size vs JSON
- Self-describing with schema ID in header
- Best for: High-throughput applications, data warehousing

**Protobuf** (Google Protocol Buffers):
- Binary serialization
- Strong typing with .proto files
- Language-agnostic (Java, Python, Go, C++, etc.)
- Efficient encoding
- Best for: Polyglot environments, gRPC integration

**JSON Schema**:
- Human-readable text format
- Easy debugging
- Widely supported
- Larger message size
- Best for: Development, debugging, REST APIs

### Compatibility Modes

| Mode | Producer Can | Consumer Can | Use Case |
|------|-------------|-------------|----------|
| **BACKWARD** | Remove fields, add optional fields | Read old data with new schema | Most common, safe for consumers |
| **FORWARD** | Add fields, remove optional fields | Read new data with old schema | Safe for producers |
| **FULL** | Add/remove optional fields only | Bi-directional compatibility | Both producers and consumers upgrade independently |
| **NONE** | Any change | Must coordinate upgrades | Development only, NOT production |
| **BACKWARD_TRANSITIVE** | BACKWARD across all versions | Read any old data | Strictest backward compatibility |
| **FORWARD_TRANSITIVE** | FORWARD across all versions | Read any new data | Strictest forward compatibility |
| **FULL_TRANSITIVE** | FULL across all versions | Complete bi-directional | Strictest overall |

**Default**: `BACKWARD` (recommended for production)

### Schema Evolution Strategies

**Adding Fields**:
```avro
// V1
{
  "type": "record",
  "name": "User",
  "fields": [
    {"name": "id", "type": "long"},
    {"name": "name", "type": "string"}
  ]
}

// V2 - BACKWARD compatible (added optional field with default)
{
  "type