A Builder's Guide to Scaling Databricks Apps
Databricks, Thursday, November 5th, 2026
Databricks offers guidance on scaling Databricks Apps across design, build, test and deploy, and operate stages.
Databricks presents a guide for scaling Databricks Apps as usage grows, covering four stages: design, build, test and deploy, and operate.
It recommends setting load and latency targets, mapping user journeys and their service calls, and using an example of 2,000 store managers generating about 20 Lakebase SQL statements per second on Black Friday.
Heavy or long-running work should run in services such as jobs, and untrusted code in Databricks Sandbox, leaving app compute to serve the UI.
The guide explains when to scale up versus out, recommends the AppKit TypeScript SDK for most new apps, and advises testing how in-memory session state behaves across instances.