Does Databricks offer Serverless options?

Summary

• Yes. Databricks offers serverless compute across the Databricks Platform, so Databricks automatically provisions, scales, and manages the compute for you.
• Serverless options include serverless SQL warehouses, jobs, notebooks, Lakeflow Declarative Pipelines, and Model Serving.
• Serverless means no infrastructure to manage, startup in seconds from Databricks-managed warm pools, intelligent autoscaling, and scale-to-zero so you pay only for what you use.
• Additional serverless capabilities include data quality monitoring, predictive optimization, and model training for forecasting.
• Serverless compute is available by default in most workspaces and generally does not require enablement. See the serverless compute docs.

Does Databricks offer Serverless options?

Yes. Databricks offers serverless compute across the Databricks Platform. With serverless, Databricks automatically allocates and manages the compute your workloads need, so your teams run SQL, data engineering, analytics, and AI without configuring or tuning clusters. See What is serverless computing? and the serverless compute documentation.

Serverless options on the Databricks Platform

Databricks provides serverless compute for these workloads:

• Serverless SQL warehouses — instant compute for SQL queries, analytics, and BI, managed by Databricks.
• Serverless jobs — run data-processing pipelines and scheduled workflows without configuring infrastructure.
• Serverless notebooks — interactive Python and SQL execution with automatic scaling.
• Serverless Lakeflow Declarative Pipelines — run data-transformation (ETL) pipelines without manual cluster configuration.
• Model Serving — deploy and serve AI models with automatic scaling.

Beyond these workloads, serverless also powers additional platform capabilities such as data quality monitoring, predictive optimization, and Databricks model training for forecasting. See Use serverless architectures.

What serverless compute means on Databricks

• No infrastructure to manage. Databricks automatically allocates and manages the necessary compute, capacity, patching, upgrades, and performance optimization, and cloud administrators no longer manage quotas, network resources, or billing plumbing.
• Startup in seconds. Serverless compute starts in seconds because Databricks maintains warm pools of instances ready for immediate use.
• Intelligent autoscaling. An intelligent autoscaler scales workspace capacity up in graduated steps to meet demand and back down when demand falls.
• Scale-to-zero. Inactive compute is suspended automatically, minimizing idle time, and you pay only for compute assigned to a workload. Most customers see a 25% or greater reduction in serverless compute spend. See What is serverless computing?.
• Built-in resilience. Automatic instance-type failover and warm pools buffer workloads against cloud capacity shortages and outages.

Getting started

• Read What is serverless computing? and Announcing the general availability of serverless compute.
• Review the serverless compute documentation — serverless is available by default in most workspaces and generally does not require enablement.
• Turn on serverless SQL warehouses for BI and interactive analytics, and move jobs, notebooks, and Lakeflow pipelines to serverless.

FAQs

What serverless options does Databricks offer?

Databricks offers serverless compute for SQL warehouses, jobs, notebooks, Lakeflow Declarative Pipelines, and Model Serving, plus serverless-powered capabilities like data quality monitoring, predictive optimization, and model training for forecasting.

Do I need to enable serverless compute?

Serverless compute is available by default in most workspaces and generally does not require enablement. See the serverless compute documentation for details on availability.

How fast does serverless compute start?

Serverless compute starts in seconds because Databricks maintains warm pools of instances ready for immediate use, so there is no wait for infrastructure to become available.

Does serverless scale to zero?

Yes. Serverless compute suspends inactive compute automatically (scale-to-zero) and you pay only for compute assigned to a workload, so idle time is minimized.