---
title: "Smarter cost controls for serverless"
description: "Databricks is extending cost control functionality to serverless compute for notebooks, jobs, and pipelines, matching the controls already available for serverless SQL and Model Serving. The new controls operate at the workload and user layer rather than the infrastructure layer, with three capabilities: attribute costs through tags on serverless compute objects, control access via entitlements on default interactive and automated compute objects, and limit spend with per-compute rate limits expressed as t-shirt sizes S, M, L, XL. Serverless Compute Access Controls are generally available, rate limits are in beta, and serverless compute tags are in beta. Monitoring via system tables and budget alerts is generally available, letting teams track consumption within hours. The full set of controls moves cost governance from machine-level caps to people- and workload-level policies, addressing runaway spend, wrong hands, and lack of accountability."
---

# Smarter cost controls for serverless

[Databricks](https://yomu.fyi/company/databricks) · Piyush Singh, Vijay Raja · Oct 29, 2026

**Type:** Announcement

## Summary

Databricks is extending cost control functionality to serverless compute for notebooks, jobs, and pipelines, matching the controls already available for serverless SQL and Model Serving. The new controls operate at the workload and user layer rather than the infrastructure layer, with three capabilities: attribute costs through tags on serverless compute objects, control access via entitlements on default interactive and automated compute objects, and limit spend with per-compute rate limits expressed as t-shirt sizes S, M, L, XL. Serverless Compute Access Controls are generally available, rate limits are in beta, and serverless compute tags are in beta. Monitoring via system tables and budget alerts is generally available, letting teams track consumption within hours. The full set of controls moves cost governance from machine-level caps to people- and workload-level policies, addressing runaway spend, wrong hands, and lack of accountability.

## Context

Databricks serverless compute usage has grown more than 3x in the past year, and thousands of customers run mission-critical notebooks, jobs, and pipelines on it. Platform teams need better visibility into serverless costs and budget control without the cluster knobs of classic compute, facing issues with runaway spend, unauthorized access, and lack of accountability.

## Approach / What changed

Three-layer cost control: attribute costs via tags on serverless compute objects (tracked in system.billing.usage); control access via entitlements on default serverless interactive and automated compute objects with Can Use and Can Manage permissions; limit spend via per-compute rate limits expressed as t-shirt sizes (S, M, L, XL) that act as autoscaling ceilings. Monitoring and budgets via system tables and account-console dashboards.

## Takeaways

- Serverless cost controls shift from machine-level caps (e.g., cluster sizes) to workload- and user-level policies such as tags, entitlements, and rate limits.
- Serverless Compute Access Controls provide two default compute objects per workspace: default serverless interactive compute and default serverless automated compute, with Can Use and Can Manage permissions.
- Rate limits on serverless compute are expressed as t-shirt sizes (S, M, L, XL) and act as an autoscaling ceiling, not a fixed cluster size; notebooks and jobs are allocated resources up to that ceiling.

## FAQ

### How do I attribute serverless costs?

Tag serverless compute objects; every DBU carries the tag to billing via system.billing.usage, which breaks out spend by user, product, and tag. A serverless\_compute\_id field further attributes spend per compute object.

### How do I control who can run serverless?

Use Serverless Compute Access Controls, which are entitlements on default compute objects. Permissions are Can Use and Can Manage. Admins remove 'All Users' and add specific groups; users outside those groups see Serverless grayed out in the compute selector.

### How do rate limits work on serverless?

Rate limits are per-compute limits expressed as t-shirt sizes (S, M, L, XL) that act as an autoscaling ceiling for notebooks and jobs. They are in Beta and can be enabled through preview settings. A shared budget for interactive sessions is coming soon.

**Tags:** [Cost Optimization](https://yomu.fyi/topic/cost-optimization), [Databricks](https://yomu.fyi/topic/databricks), [Serverless](https://yomu.fyi/topic/serverless)

- Source: [Databricks](https://www.databricks.com/blog/smarter-cost-controls-serverless)
- Source URL: https://www.databricks.com/blog/smarter-cost-controls-serverless
- Ingested by Yomu: 2026-10-29T18:01:08.333Z

[Read original post](https://www.databricks.com/blog/smarter-cost-controls-serverless)
