specification: FinOps Framework
specificationVersion: '1.0'
schema: https://www.finops.org/framework/
provider: Azure Databricks
providerId: azure-databricks
created: '2026-05-04'
# Provenance stamped 2026-08-11: this artifact was written by the API Evangelist
# bulk sweep dated 2026-05-04, not harvested from the provider. See roadmap#35.
method: generated
modified: '2026-05-05'
reconciled: true
tags:
  - Analytics
  - Lakehouse
  - Spark
  - FinOps
  - FOCUS
description: 'FOCUS-aligned FinOps for Azure Databricks: dual-meter pricing where DBUs (Databricks Units)
  are billed by Databricks via Azure Marketplace and the underlying Azure VM, storage, and network are
  billed by Azure. Reserved DBUs (DBCU) and Azure Reservations stack for committed-use savings.'
sources:
  - https://azure.microsoft.com/en-us/pricing/details/databricks/
  - https://learn.microsoft.com/en-us/azure/databricks/admin/account-settings/usage
alignedWith:
  framework: FinOps Foundation Framework
  frameworkUrl: https://www.finops.org/framework/
  dataSpec: FOCUS
  dataSpecVersion: '1.3'
  dataSpecUrl: https://focus.finops.org/focus-specification/v1-3/
publisherName: Microsoft Corporation
serviceCategory: Data Analytics
billingModel:
  pricingCategory: Pay-As-You-Go + Committed Use
  billingFrequency: Monthly
  billingCurrency: USD
  chargeCategories:
    - Usage
    - Purchase
    - Tax
    - Adjustment
focusColumns:
  ServiceName: Azure Databricks
  ServiceCategory: Data Analytics
  ProviderName: Microsoft
  PublisherName: Microsoft Corporation
  InvoiceIssuerName: Microsoft Corporation
  PricingCategory: Pay-As-You-Go
  BillingCurrency: USD
  ChargeCategory: Usage
meters:
  - name: dbu_consumption
    description: Databricks Unit consumption (per second) by workload type
    unit: DBU
    aggregation: sum
    dimensions:
      - workload_type
      - tier
      - region
      - workspace
      - photon
  - name: vm_compute_hours
    description: Underlying Azure VM time (billed separately by Azure)
    unit: instance-hour
    aggregation: sum
    dimensions:
      - vm_sku
      - region
      - workspace
  - name: serverless_compute
    description: Serverless SQL / Model Serving / Jobs compute charged as DBUs without separate VM line
    unit: DBU
    aggregation: sum
    dimensions:
      - service
      - region
  - name: storage_capacity
    description: Underlying ADLS Gen2 / managed storage
    unit: GB-month
    aggregation: avg
    dimensions:
      - region
  - name: model_serving_queries
    description: Inference queries against Model Serving endpoints
    unit: query
    aggregation: sum
    dimensions:
      - endpoint
  - name: vector_search_capacity
    description: Vector Search endpoint storage and throughput units
    unit: endpoint-hour
    aggregation: sum
principles:
  - name: Visibility
    description: Use the Azure Databricks usage page (system tables) and Azure Cost Management to view
      DBU consumption by workspace, cluster, job, and tag; integrate with FOCUS exports.
  - name: Allocation
    description: Apply cluster tags and pool tags propagated to Azure VM cost; cross-reference with system.billing.usage
      for chargeback by job, user, or workspace.
  - name: Optimization
    description: Use Photon for SQL/ETL; choose Jobs Compute over All-Purpose for production workloads;
      enable autoscaling and auto-termination; right-size VM SKUs; commit DBCUs and Azure Reservations
      for predictable usage.
  - name: Accountability
    description: Set workspace-level budgets in Azure Cost Management; assign workspace admins as owners;
      review high-cost queries and idle clusters in monthly FinOps reviews.
maintainers:
  - FN: Kin Lane
    email: kin@apievangelist.com