System tables
System tables is a Databricks data governance capability, introduced June 2023.
Billable usage log delivery System tables
A Databricks-hosted analytical store of an account's own operational data - billing, access, compute, lineage and more - exposed as read-only Unity Catalog tables in the system catalog.
- Your own platform becomes a dataset you can join: the billing, audit and lineage data behind the console arrives as ordinary tables you query with SQL.
- They are read-only, so this is the rare Unity Catalog schema where your write privileges mean nothing at all.
- The platform went generally available before most of its own tables did - the September 2024 GA covered the system tables platform plus exactly two billing tables.
Limitations: Read-only and unmodifiable, they must be enabled from a Unity Catalog-enabled workspace before anything appears, and free retention varies per table unless an account admin opts into the configurable-retention Beta (September 2026), which sets a flat 395 days for every supported table and lets admins choose 30 to 3,650 days, though data kept past 395 days will be billed once the feature goes GA and the system.data_classification and system.data_quality_monitoring schemas are excluded. Since August 26, 2026 the SQL text in system.query.history reads <REDACTED> by default - only account admins and members of the databricks_pii_access group see the full statement.
Open in REbricked →- Category
- Data governance
- Also known as
- system catalog, system.billing
- Verified
- 2026-09-20
Sources
- Official Official Databricks / Microsoft docs
- Official Databricks docs: query history system table - statement text is <REDACTED> by default since August 26, 2026
- Official Databricks release notes: configurable retention for system tables in Beta (September 9, 2026)
- Official Databricks release notes: query text masked by default in system tables and APIs (August 26, 2026)
- Official Databricks docs: pipeline events system table (Beta since August 31, 2026)