Redpanda Connect now includes a native Iceberg output component that writes streaming data directly to Apache Iceberg tables via a declarative YAML pipeline. It supports 300+ input sources (HTTP webhooks, Postgres CDC, GCP Pub/Sub, etc.), inline data transformation using Bloblang, dynamic multi-table routing via interpolation, and automatic schema evolution from raw JSON without a schema registry. The connector speaks the Iceberg REST Catalog API, making it compatible with Polaris, AWS Glue, Databricks Unity Catalog, Snowflake Open Catalog, and GCP BigLake. It uses data-driven flushing to avoid the small file problem and runs with as little as 0.1 vCPU. Available in Redpanda Connect v4.80.0 as an enterprise-tier connector, with append-only ingestion initially and upserts on the roadmap.
Table of contents
From stream to table, without the detours #Example pipeline #Reshape your data before it lands #Works with your catalog #Small in size. Big on benefits #When to use Iceberg output #Getting started #Sort: