Observe on Apache Iceberg: Unlocking Open Observability

Snowflake's Observe product now supports storing telemetry data as Apache Iceberg tables in customer-owned Amazon S3 buckets, currently in private preview. Observability data (logs, metrics, traces) is written as standard Iceberg v3 tables, enabling any Iceberg-compatible engine — Spark, DuckDB, Trino, PyIceberg — to query it directly via a read-only Iceberg REST Catalog. The full Observe experience (dashboards, monitors, AI SRE, OPAL transformations) works identically on Iceberg-backed data. Key benefits include data portability and ownership, cost reduction by decoupling storage from compute, elimination of ETL pipelines to join telemetry with business data, and compliance-friendly direct access to raw telemetry from customer-controlled storage.