First public release (TinybirdJsonToStruct SMT)

What’s included

• Kafka Connect SMT that converts schemaless JSON Map values (JsonConverter with schemas.enable=false) into a typed Struct with an explicit Connect Schema.
• Enables JDBC Sink (and other schema-aware sinks) to write proper SQL types instead of generic JSON blobs.

Configuration

• schema.name (default: TinybirdRecord) — schema name for the generated Struct.
• fields (required) — field mapping: fieldName:type,fieldName2:type2,...
  • Supported types: int64, int32, string, boolean, float64, float32, timestamp, array_string
• pk.field (optional) — adds a static primary key field to the value schema.
• pk.value (optional) — static value for pk.field.

Timestamp handling

• timestamp fields support:
  • epoch millis (number or numeric string)
  • ISO-8601 (e.g. 2024-11-22T00:43:33Z)
  • yyyy-MM-dd HH:mm:ss.SSS (UTC)
• Special case: if a field named updatedAt is configured as timestamp, it is populated from the Kafka Connect record timestamp (fallback: current time).

Artifacts

• tinybird-json-to-struct-<version>.zip — contains the compiled SMT JAR ready to be dropped into a Kafka Connect plugin path.

Notes

• This is the first public release. APIs and configuration may evolve based on early feedback.
• Please report issues or suggestions via GitHub Issues.