Observe on Apache Iceberg: Unlocking Open Observability
Snowflake’s Observe platform now supports Apache Iceberg tables in private preview, allowing users to store observability data in open formats within their own Amazon S3 buckets while maintaining full functionality.
Snowflake’s Observe platform has introduced a private preview of Observe on Apache Iceberg, enabling users to store observability data as Iceberg tables directly in their Amazon S3 buckets. This approach leverages open standards, allowing data to be accessed and queried by any Iceberg-compatible engine while remaining under the user’s control and governance policies. The move addresses longstanding challenges in observability data management, where vendor-specific formats and siloed storage have historically hindered cross-team collaboration and data integration.
The integration allows observability data—logs, metrics, and traces—to be ingested via the existing Observe pipeline and written as standard Iceberg tables in the user’s S3 bucket. Users retain full ownership of the data, including the ability to revoke Observe’s access without losing the dataset. Snowflake-managed Iceberg tables (v3 specification) handle automatic compaction, file sizing, and metadata upkeep, ensuring high-performance queries without manual maintenance. The v3 spec also introduces richer data types and performance improvements, enhancing usability for downstream analytics.
Observe on Iceberg maintains the full Observe experience, including monitoring, dashboards, and AI-assisted root cause analysis, while enabling interoperability with external tools. Users can query observability data alongside business data in a single analysis, run batch transformations, audit raw telemetry for compliance, or train ML models directly from the stored data. This eliminates the need for separate pipelines or data replication, reducing costs and operational overhead while dissolving traditional data silos between engineering and analytics teams.
The solution decouples storage from compute, allowing users to retain large volumes of telemetry at lower costs compared to proprietary observability platforms. Storage occurs in the user’s S3 bucket, with query costs incurred only when accessing the data. Observe on Iceberg is available in private preview, with broader availability planned. Interested users are encouraged to apply for participation in the preview program.