Databricks Cost Estimation Template

[Phani1]

Hi Databricks Team,

Is there a standard Databricks cost estimation template(xl), sizing calculator, or TCO tool that allows us to provide the following inputs and derive an approximate monthly and annual platform cost:

• Source systems and their types (e.g., SQL Server, AS400, JDE, Maximo)
• Batch vs. streaming workload split
• Total data volume (historical migration + daily/monthly incremental)
• Number of ETL jobs (e.g., 600+ SSIS packages) and their run frequency (hourly, daily, weekly)
• Average job execution time and SLA requirements
• Reporting workload — number of active reports, concurrent users, and query patterns
• AI/ML and GenAI use cases — model training, serving, and inference needs
• Number of environments (Dev, QA, Prod)

This would help us quickly estimate the expected DBU consumption and total cost without having to manually calculate each workload individually.

[emma_s · Accepted solution]

Hi, There isn't anything publicly available that I'm aware of. For this kind of complex migration I'd recommend working with your account team. As somebody who does Databricks sizing a lot, it's a nuanced art which I suspect is why we don't have any rough calculator. The solutions architect on your account will have access to the latest and greatest ways of sizing though.

Thanks,

Emma

[Phani1]

Thank you!!

[robert7878]

A standardized cost estimation template would be extremely useful for projects of this scale. Combining workload frequency, execution time, data volume, concurrency, and environment requirements into one model would make monthly and annual cost projections much easier to compare. I’d also recommend including proportional calculations when estimating resource consumption across different workloads; a Dreisatz Rechner Online can be helpful for checking those figures and validating assumptions.

[Harshad_B]

@Phani1 @robert7878 - You can build a AI/BI dashboard pretty easily using Genie, just give what is required and how you want it to be represented.