Databricks serverless, cost

Irene Moser 20 Reputation points
2026-08-16T01:11:55.2066667+00:00

I read this page https://azure.microsoft.com/en-us/pricing/details/databricks/#pricing

and am none the wiser. What is a DBU with serverless? Do I get charged when the Databricks instance exists or when I run code on it?

Cost Management
Cost Management

A Microsoft offering that enables tracking of cloud usage and expenditures for Azure and other cloud providers.

0 comments No comments

Answer accepted by question author
Marcin Policht 107.5K Reputation points MVP Volunteer Moderator
2026-08-16T03:41:50.84+00:00

A DBU (Databricks Unit) with serverless is a single, all-inclusive billing unit that measures the processing power your workloads consume, bundling both the Databricks software fee and the underlying cloud infrastructure costs into one rate.

You are charged when you run code or execute queries on active serverless compute. Billing begins when your query or serverless job starts processing data, and usage is measured on a per-second basis.

Unlike traditional classic clusters, where you can incur cloud infrastructure costs simply because a cluster remains running, serverless resources automatically suspend when they are no longer being used. Therefore, you do not continue accumulating DBU charges while the serverless resource is suspended or shut down.

If a serverless SQL warehouse or notebook environment is idle, it will automatically suspend according to its configured settings. Once suspended, it is no longer consuming serverless compute, so no additional DBU charges accumulate during that suspended period.


If the above response helps answer your question, remember to "Accept Answer" so that others in the community facing similar issues can easily find the solution. Your contribution is highly appreciated.

hth

Marcin

Was this answer helpful?

2 people found this answer helpful.

0 additional answers

Sort by: Most helpful

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.