Calculator inputs
Use the provisioned model for nodes and reserved planning. Use the serverless model for RPU-based billing.
Formula used
Provisioned compute = node hourly rate × node count × active hours × AZ factor.
Reserved adjustment = provisioned compute − discount + monthly upfront amortization.
Concurrency scaling = billable overage seconds × cluster on-demand rate ÷ 3600.
Serverless compute = average RPUs × billable seconds ÷ 3600 × RPU-hour rate.
Serverless reservation savings = compute × coverage % × discount %.
Storage = GB-month × storage rate. Spectrum = TB scanned × scan rate.
Total monthly estimate = compute + storage + snapshots + scans + transfer + support + other costs − eligible credits.
How to use this calculator
- Select provisioned or serverless based on your Redshift deployment.
- Enter your region label and confirm every editable default rate.
- For provisioned clusters, add node count, active hours, and any reservation effects.
- For serverless, estimate sessions, average runtime, average RPUs, and reservation coverage.
- Enter managed storage, snapshot storage, spectrum scans, and transfer usage.
- Click calculate to see the result above the form.
- Use the CSV and PDF buttons to export the estimate summary.
Example data table
| Scenario | Model | Key usage | Monthly estimate idea |
|---|---|---|---|
| Development analytics | Provisioned | 2 nodes, 240 active hours, 2 TB storage | Useful when clusters pause outside working hours. |
| Production warehouse | Provisioned Multi-AZ | 4 nodes, 730 hours, 40 TB storage | Good for resilience planning and reservation comparison. |
| Intermittent BI workload | Serverless | 80 sessions, 240 seconds, 48 average RPUs | Best for bursty usage without fixed cluster sizing. |
| Lake-heavy reporting | Serverless | 12 TB external scans plus 8 TB storage | Highlights compute and external scan impact together. |
Frequently asked questions
1. What does this calculator estimate?
It estimates monthly and annual Redshift spending using provisioned nodes or serverless RPUs, plus storage, snapshots, scans, transfer, support, and custom charges.
2. Why are the rate fields editable?
AWS prices can differ by region, node type, purchase option, and date. Editable fields let you adjust this page quickly without rewriting the calculator.
3. How is serverless billing modeled?
The page multiplies average RPUs by billable seconds and the RPU-hour rate. It also respects the 60-second minimum by comparing runtime against session minimums.
4. Does the tool handle reserved pricing?
Yes. For provisioned clusters, you can apply a custom discount and monthly amortized upfront amount. For serverless, you can model reservation coverage and discount.
5. Are concurrency and Spectrum charges separate?
Provisioned estimates treat concurrency scaling overages separately. Spectrum scanning is also separated. In serverless mode, data-lake query compute is already included in serverless usage.
6. Can I model free credits?
Yes. Serverless mode includes a field for trial credit. Provisioned mode includes free concurrency seconds so daily earned credits can offset burst costs.
7. Is the estimate exact enough for invoices?
No. It is a planning tool. Final billing still depends on real AWS metering, regional rates, reservation contracts, and workload behavior.
8. What should I update first?
Start with deployment model, node or RPU rate, active usage, managed storage, and scans. Those usually drive the biggest changes in cost.