Databricks: Scoped Freshness and Volume Collection

For new Databricks integrations, Monte Carlo now limits freshness and volume collection to tables that have a monitor, rather than every table ingested. Because Databricks requires a DESCRIBE call per table run serially, scoping collection to monitored tables keeps that work targeted and reduces the time clusters stay active.

This is on by default for new integrations; existing integrations are unchanged and continue collecting across all tables. Either behavior can be switched through our Support agents. When scoped collection is active, a table's collection begins when its first monitor is added, so it will have no history initially and needs a brief warm-up period before alerting. For teams who prefer broader coverage from day one, full-collection mode remains available, with ingestion controls to block expensive calls at the schema or table level if needed.