ADF costing logic

Hi @HimanshuSinha

The way ADF billing works is that you’re effectively paying for each of the “moving parts” you use in your pipelines. At a high-level you’ll see charges in these buckets:

1. Orchestration (pipeline activities)
   • You pay per activity run. Every time an activity (Copy, Lookup, Execute Pipeline, etc.) executes, it’s counted toward your orchestration cost.
   1. Data movement (Copy activity)
      • If you’re copying data between stores using the Azure Integration Runtime, you’re charged per “Data Integration Unit”-hour (DIU-hr). The more parallelism you ask for, the more DIUs you consume.
      1. Mapping Data Flows
         • Data flows run on Azure Databricks/Spark under the covers. You’re charged per vCore-hour (or Debug hour if you leave debug turned on).
         1. SSIS Integration Runtime (if you’re running SSIS packages)
            • You pay for the IR compute nodes you spin up, similar to how you’d pay for any Azure VM.
            1. Triggers, external activities, monitoring, etc.
               • Some small charges may accrue for things like event-based triggers or monitoring pipeline runs via Azure Monitor.

Because rates vary by region and by your chosen level of parallelism/compute, your total bill will depend on:

• Which activities you’re using and how many times they run
• Volume of data you’re moving and the DIUs you allocate
• How long your Data Flow or SSIS runtime clusters stay up
• How often you-re triggering pipelines

Hope that gives you the “big picture.”

If you can share a few more details about your scenario, I can help nail down a rough monthly estimate:
• Which region(s) are you running in?
• How many pipelines/activities run per day?
• Rough data volume (GB) in your Copy activities and DIUs you’ve allocated
• Any Mapping Data Flows—how many vCore-hours do those consume?
• Are you using a self-hosted IR or Azure-hosted IR?

Reference docs:
• What is Azure Data Factory? Introduction + How it works (pipelines, mapping data flows)
https://learn.microsoft.com/azure/data-factory/introduction
• Azure Data Factory overview (control flow, SSIS, Spark, scheduling, triggers)
https://learn.microsoft.com/rest/api/datafactory/
• Features of Azure Data Factory (connectivity, triggers, data preview, security)
https://learn.microsoft.com/azure/data-factory/introduction#features-of-azure-data-factory
• Branching and chaining activities (example of orchestrating a pipeline)
https://learn.microsoft.com/azure/data-factory/tutorial-control-flow#create-visual-studio-project

Hello !

Thank you for posting on Microsoft Learn Q&A.

ADF pricing is consumption based and in my opinion there is no single flat cost for having ADF.

You’re charged based on what your pipelines actually do but I can tell you that the core billing meters are orchestration activity runs, data movement for copy activity and vCore-hours for mapping data flows.

You can also find in the doc pipeline activity execution, external activity execution, artifact operations and SSIS IR runtime that all can be added to the bill.

https://learn.microsoft.com/en-us/azure/data-factory/plan-manage-costs

A simple way to think about it is total ADF cost = orchestration runs + copy DIU hours + data flow vCore-hours + any extra IR or operations or SSIS IR charges. This is why 2 factories with the same number of pipelines can cost very different amounts depending on frequency, runtime duration, data volume and whether they use copy only or heavy mapping data flows.

If you are seeking an estimation, you can use the ADF pricing calculator then validate with a small test workload and check actual consumption in cost analysis and the ADF Monitor experience at pipeline run level.