Why Did My Databricks Bill Go Up?

Databricks bills increase when DBU consumption, cloud infrastructure, serverless usage, storage, or data transfer increases.

Start with system.billing.usage and group by SKU, workspace, cluster, job, and tag. Compare the first abnormal day with cluster runtime, job frequency, worker count, and request volume. A change in SKU can indicate a new compute type or product.

Common causes include interactive clusters left running, a job scheduled more often, larger workers, new serverless workloads, Photon or AI features, and increased data volume. The underlying notebook or SQL may be unchanged while its schedule or input grows.

Compare the affected period with a representative baseline and the cloud-provider invoice. Record the first date of the increase and the configuration change that explains it.