Compute nodes plus managed storage for a Redshift data warehouse.
A provisioned Redshift cluster charges for every hour it exists, busy or not. Right-size the node count, pause off-hours, and compare with other warehouses on the BigQuery cost calculator and Snowflake credit calculator.
The Amazon Redshift Cost Calculator estimates your monthly bill for running a provisioned Redshift data warehouse. It multiplies your number of compute nodes by the hourly price per node and the hours per month the cluster runs, then adds the cost of managed storage based on the gigabytes you keep. The result is the total monthly figure combining compute and storage. Compute is usually the largest driver, since node count, node type price, and uptime compound together β a cluster left running around the clock costs far more than one that runs only during business hours or scheduled ETL windows.
The main trade-off to watch is the balance between node count and runtime. Adding nodes speeds up queries but raises the hourly cost linearly, so scaling up a cluster that runs continuously multiplies your bill quickly. If your workload is intermittent, pausing the cluster during idle hours or resizing to fewer nodes can cut compute spend substantially, while storage stays relatively cheap and predictable. Adjust the hours-per-month input to model pausing, and change the node count to compare configurations. Seeing these numbers before you commit helps you right-size the warehouse rather than over-provisioning and paying for capacity that sits idle most of the day.
RA3 nodes bill per node-hour for compute, with managed storage charged separately per GB-month (about $0.024). Older DC2 nodes bundle storage into the node price. Redshift Serverless instead bills per RPU-second of actual query time.
If the cluster only runs queries part of the day, a provisioned cluster still bills every idle hour. Pausing it off-hours, or switching to Serverless so you only pay while queries run, is often much cheaper for spiky workloads.