Serverless rate limits

Caps how far an automated serverless workload, a job or a Spark Declarative Pipeline, can autoscale. A workspace admin sets a size on a custom automated serverless compute object, from Small (about 60 DBUs per hour) to 2X-Large (about 960), and each job or pipeline that uses it is capped on its own. Interactive notebooks and Databricks Connect are not rate-limited, and Databricks Runtime 17.3.1 or above is required. Databricks has said shared budgets for interactive serverless work are coming soon.

Status and dates are extracted from public sources only. We don't see the private AE-shared roadmap. If something looks wrong, tell us.