AWS Redshift Serverless cost: you pay per RPU-hour of query time
Redshift Serverless bills per Redshift Processing Unit (RPU) hour while queries run, plus managed storage, with no cluster to pause. It suits variable, intermittent analytics; steady heavy load is cheaper on a provisioned cluster. Here is how to size it.
Quick answer
Redshift Serverless bills per RPU-hour (Redshift Processing Unit) while queries run, metered per second with a minimum, plus managed storage per GB-month. It scales capacity to the workload with nothing to pause. Serverless suits variable, intermittent analytics; a provisioned cluster (with reservations) is cheaper for steady, heavy query load.
Redshift Serverless removes cluster management from data warehousing: instead of provisioning nodes, you pay for the compute your queries consume, measured in RPUs. The convenience is priced, so the decision against a provisioned cluster is the familiar variable-versus-steady tradeoff.
RPU-hours and storage
You set a base capacity in RPUs and pay per RPU-hour, metered per second with a minimum charge per query burst, only while queries run. Idle periods cost nothing for compute. Storage is managed and billed per GB-month separately. So a warehouse queried in bursts pays for those bursts, not for a cluster sitting idle overnight.
Serverless vs provisioned
| Mode | Bills for | Best when |
|---|---|---|
| Serverless | RPU-hours while querying + storage | Variable, intermittent, spiky |
| Provisioned | Node-hours, always on | Steady, heavy, predictable |
Serverless wins when query load is variable or the warehouse is often idle, since you avoid paying for provisioned nodes doing nothing. A provisioned cluster, especially with reserved nodes, is cheaper for steady, heavy, predictable query volume, the same crossover as Redshift versus Athena.
Controlling serverless cost
Set a sensible base RPU capacity and a maximum so a runaway query cannot scale cost without bound, and use the usage limits and alerts to cap spend. Reduce RPU-hours the same way you reduce any query cost: good table design, sort and distribution keys, and avoiding scans of data you do not need. If a serverless workload grows into steady heavy use, re-evaluate against a provisioned cluster.
FAQ
How is Redshift Serverless priced?
Per RPU-hour (Redshift Processing Unit) while queries run, metered per second with a minimum, plus managed storage per GB-month. Idle periods cost nothing for compute, so a warehouse queried in bursts pays for those bursts rather than an always-on cluster.
Is Redshift Serverless cheaper than a provisioned cluster?
For variable, intermittent, or often-idle query load, yes, because you avoid paying for provisioned nodes sitting idle. For steady, heavy, predictable load a provisioned cluster, especially with reserved nodes, is cheaper. Match the mode to your query pattern.
What is an RPU?
A Redshift Processing Unit, the compute-billing unit for Redshift Serverless. You set a base RPU capacity and pay per RPU-hour while queries run. More RPUs mean more parallelism and cost, so sizing the base and maximum controls spend.
How do I cap Redshift Serverless cost?
Set a base RPU capacity and a maximum so a runaway query cannot scale cost without bound, and configure usage limits and alerts to cap spend. Reduce RPU-hours with good table design, sort and distribution keys, and avoiding unnecessary scans.
When should I switch to a provisioned Redshift cluster?
When query load becomes steady and heavy enough that always-on provisioned nodes, especially reserved, cost less than the RPU-hours a serverless workload of the same volume would consume. As load grows and stabilizes, provisioned becomes cheaper.
Does C3X estimate Redshift cost?
C3X prices provisioned Redshift clusters from your Terraform directly. For Serverless, cost depends on RPU-hours consumed by queries, which is a usage input you model, plus managed storage that scales with data volume.
What to do next
Compare serverless and provisioned before you choose. C3X reads your Terraform and prices Redshift and the rest of your stack against a live catalog. Start with the quickstart.
Pricing references
Related posts
- AWS Athena cost optimization: you pay per terabyte scanned
Athena charges $5 per TB of data scanned, so cost is decided by how much data each query reads, not how often you query. Partitioning, columnar formats, and compression cut scans by an order of magnitude. Here is how.
- Redshift vs Athena cost: provisioned cluster or pay-per-scan?
Athena bills $5/TB scanned and costs nothing idle; Redshift bills for a provisioned cluster (~$793/month per node) or per RPU. Here's the crossover by query volume and how table design cuts both.
- AWS AppSync cost: per query, per real-time message, per connection
AppSync bills GraphQL query and data-modification operations per million, real-time subscription messages per million, and connection-minutes. Chatty clients and broad subscriptions drive the bill. Here is how the managed-GraphQL pricing works.
- AWS Transfer Family cost: per-hour endpoints plus data transferred
Transfer Family bills about $0.30 per hour for each enabled protocol endpoint (SFTP, FTPS, FTP) plus about $0.04 per GB uploaded and downloaded. The always-on endpoint fee is the surprise for low-volume use. Here is the breakdown.
Share this post
Try C3X on your own Terraform
Free and open source. No API key required. One command to install, one command to estimate.