Azure Databricks Cost Structure
Azure Databricks workspaces can be deployed with one of three price tiers: standard, premium, or trial. The primary difference between the standard and premium price tiers is that role-based access control for workspace objects and Azure AD credential passthrough is only available with the premium price tier. The trial price tier is a 14-day free trial of the Azure Databricks premium price tier.
Pricing for Spark clusters created in Azure Databricks consists of two primary components: the cost of the driver and worker node VMs and the processing cost. Processing cost is measured by the number of Databricks Units (DBUs) consumed during cluster runtime. A DBU is a unit of processing capability per hour, billed on per-second usage. You can easily calculate the number of DBUs usage by multiplying the total number of cluster nodes by the number of hours the cluster was running. For example, a cluster with 1 driver node and 3 worker nodes that ran for a total of 2 hours consumed 8 DBUs (that is, 4 nodes × 2 cluster runtime hours).
While the Azure VM cost will remain the same regardless of which price tier the Azure Databricks workspace was deployed with, the DBU price will vary. Table 5.1 lists the DBU price differences for the standard and premium price tiers.
TABLE 5.1 Standard and premium tier DBU prices
Workload | Standard Tier DBU Price | Premium Tier DBU Price |
All-Purpose Compute | $0.40 DBU/hour | $0.55 DBU/hour |
Jobs Compute | $0.15 DBU/hour | $0.30 DBU/hour |
Jobs Light Compute | $0.07 DBU/hour | $0.22 DBU/hour |
Leave a Reply