Fivetran is a fully managed ELT platform for moving data from SaaS applications, operational databases, and event platforms into cloud warehouses and lakes. Its Managed Data Lake Service writes source data as Parquet in Apache Iceberg or Delta Lake tables, reducing the operational burden of schema mapping, metadata coordination, and table lifecycle management. section 6.4.4
Fivetran’s Iceberg integration provides:
- Cloud storage targets across Amazon S3, Azure Data Lake Storage, and Google Cloud Storage, with GCS support identified as beta. section 6.4.4
- A Fivetran Iceberg REST Catalog that maintains table metadata and supports engines implementing Iceberg’s REST interface. section 6.4.4
- Automated compaction, old-snapshot cleanup, orphan-file cleanup, stale-metadata cleanup, and column-level minimum and maximum statistics. section 6.4.4
- Compatibility with Spark, Dremio, Athena, and Snowflake, plus automatic renaming for reserved Iceberg field names. section 6.4.4
Managed source-to-Iceberg flow
Fivetran pulls data from Salesforce or PostgreSQL, writes Parquet files in Iceberg format to Amazon S3 through its REST Catalog, and makes the resulting tables available to Snowflake or Dremio. A bank can evaluate Fivetran alongside Airbyte and Qlik when onboarding APIs, SaaS applications, and operational databases. section 6.4.4 section 4.5.1
