Rubrik extends data protection to Apache Iceberg tables on AWS

Rubrik said the feature targets a gap in native Iceberg tooling. Iceberg snapshots are metadata pointers rather than copies, so a table deleted, encrypted by ransomware or overwritten by a malfunctioning artificial intelligence agent loses its snapshot history. Restoring the underlying files does not rebuild the catalog entry, forcing data teams to manually reassemble table metadata before any query can run again.

Apache Iceberg Protection copies the metadata, not only the underlying data files. Recovery re-registers the table in the catalog, letting analysts query the restored table in Amazon Athena, Apache Spark or Trino immediately. If only files are restored, the result is a set of Parquet files requiring manual repair, Rubrik said.

The capability covers Iceberg tables in the AWS Glue Data Catalog and Amazon S3 tables. Backups are immutable wherever they are stored, whether in the customer’s own AWS account or in the air-gapped Rubrik Cloud Vault. Rubrik said it is the only Iceberg-aware protection offering with the in-account option.

After an initial full copy, Rubrik captures only the most recent compacted snapshots. The company said this approach keeps backup windows manageable once tables reach petabyte scale. Customers can choose the S3 storage class, so rarely accessed backup data does not have to sit in a premium tier. An on-demand backup can also be taken ahead of risky work such as a schema change or a large batch write.

Restores can land on an isolated Iceberg branch so teams can inspect the data and schema before promoting the table back into production.

“Data lakes have grown beyond analytics and AI to now run the entire business, and Apache Iceberg is one of the most critical systems built on them,” said Anneka Gupta, chief product officer at Rubrik. Gupta said the company has long protected the object storage underneath and now covers the lakehouse itself, with the same policies reaching the databases, pipelines and code around it.

Apache Iceberg Protection is scheduled to be generally available later this month. Rubrik has been adding cloud-native workload protections, including Cloud SQL protection and governance for Gemini agents at Google Cloud Next in April and Agent Cloud, a control layer for enterprise AI agents, in October.