Instant compute with minimal management and lower cost
With Databricks Serverless SQL, administrators don’t need to configure or manage clusters, and users don’t need to wait for clusters to start up or scale out when running queries. With an active server fleet managed by Databricks, new compute capacity is typically available in about 15 seconds. The best part? You only pay for Serverless SQL when users start running queries.
The public preview of Serverless SQL is available on AWS. Please sign up here to request access.
Instant compute: Users can get to work immediately, as new compute capacity is typically ready in about 15 seconds. No waiting for several minutes as clusters start up or scale out. When users are done, Databricks will wait 10 minutes (by default) then shut down the cluster.
Minimal management: It only takes one click to enable Serverless SQL. There is no configuration or performance tuning, and patching and upgrades are managed automatically.
Lower cost: Serverless SQL optimizes compute to match your usage, avoiding cost from over-provisioning and idle capacity when users are not active.