Dataproc
GCP Dataproc — commitment brief
Data sourced: March 2026. Verify current figures at the Google Cloud Pricing Calculator.
Coverage summary
Dataproc has two deployment modes with distinct CUD coverage: Dataproc cluster mode and Dataproc Serverless. Dataproc cluster mode runs jobs on Compute Engine VMs and does not have its own CUD product — coverage comes entirely through Compute Engine resource-based CUDs or compute flexible CUDs applied to the underlying VM instances. Dataproc Serverless (formerly Spark Serverless) is a managed execution environment where workloads run without user-managed VMs; it participates in the spend-based CUD program (BigQuery CUD scope) and is explicitly listed as a covered service.
What is covered: For Dataproc cluster mode — the Compute Engine VM costs for master and worker nodes are covered by compute flexible CUDs (28% at 1-yr, 46% at 3-yr) or resource-based CUDs. For Dataproc Serverless — compute execution costs are covered under the BigQuery spend-based CUD umbrella.
What is not covered: The Dataproc management fee (per-vCPU-hour charge on top of Compute Engine pricing) is not covered by CUDs. Persistent Disk storage for HDFS and cluster disks, Cloud Storage costs, network egress, and Dataproc cluster VMs running as Spot/Preemptible are excluded from CUDs.
CUD types
Dataproc cluster mode: no Dataproc-specific CUD. Coverage flows through Compute Engine compute flexible CUDs (28%/1-yr, 46%/3-yr) applied to the underlying VMs. Dataproc Serverless: covered under BigQuery spend-based CUDs (10%/1-yr, 20%/3-yr discount on eligible spend). Note: resource-based Compute Engine CUDs can also cover Dataproc cluster VMs if you have existing resource commitments for the same machine family and region.
Machine type / SKU coverage
Dataproc cluster mode — General Purpose VMs (N1, N2, E2, etc.)
Compute Flex CUD
~28%
~46%
Applies to underlying Compute Engine VM costs
Dataproc cluster mode — HPC Optimized VMs (H3, H4D)
Compute Flex CUD
~17%
~38%
H3/H4D flex CUD: 17% (1-yr), 38% (3-yr)
Dataproc cluster mode — Memory Optimized VMs (M1, M2, M3)
Compute Flex CUD
no flex discount
~63%
M-series flex CUD: no 1-yr discount; 3-yr only at 63%
Dataproc cluster mode — Accelerator VMs (A2, A3, G2)
Resource CUD only
~37%
~55%
Not covered by flex CUD; resource CUD required
Dataproc cluster mode — Spot / Preemptible VMs
❌ No
—
—
Spot pricing; no CUD coverage
Dataproc management fee (per-vCPU-hr)
❌ No
—
—
Separate Dataproc SKU; not covered by any CUD
Dataproc Serverless (Spark Serverless)
BigQuery Spend CUD
~10%
~20%
Covered under BigQuery CUD spend umbrella
HDFS / Persistent Disk storage
❌ No
—
—
Storage not covered
Cloud Storage
❌ No
—
—
Not covered
Regional availability
Available in all commercial GCP regions. Dataproc's regional availability follows Compute Engine — when new Compute Engine regions are added, they become available for Dataproc automatically. CUD coverage for Dataproc cluster-mode workloads is subject to the regional restrictions of the underlying VM families: GPU-backed clusters (A2, A3, G2) are limited to the zones where those machine types are available, and M-series memory-optimized nodes are only available in select regions.
⚠️ CUD availability varies by machine type and region. Always verify at GCP regions and zones before purchasing.
Archera
Google Cloud Dataproc is within Archera's commitment management scope for cluster-mode workloads. Archera manages Dataproc cluster coverage through Compute Engine compute flexible CUDs — sizing commitments across your combined Dataproc, GKE, and Compute Engine footprint. For Dataproc Serverless, coverage is managed through BigQuery spend-based CUDs. Archera can wrap all commitments in a GRI/GSP to eliminate downside risk on over-commitment.
Sources
⚠️ Discount percentages are approximate. Dataproc cluster-mode CUD coverage depends on the underlying VM family and your existing Compute Engine commitments. Always verify with the Google Cloud Pricing Calculator.
Last updated
Was this helpful?