Pricing calculator
Describe what you run and see what it costs. Compute bills per second at the same rates on every plan, so your plan is set by what you need to run, not by how much of it you run.
//Start from
//Workloads
Single node queries
$51.00 / mo
oleander sizes this to 4 vCPU, 8 GB, at $6.12 per hour of query time. $0.0085 per run.
Spark jobs
$9.04 / mo
1 driver and 2 executors: 5 vCPU, 20 GB $0.08 per run.
Iceberg maintenance
$6.21 / mo
Runs on Spark once a day per table, compacting files and expiring snapshots. 1 driver and 4 executors: 18 vCPU, 36 GB. $0.07 per run.
//Storage
Stored in
//How this is estimated
- Single node compute costs $0.000425 per vCPU second of active query time. oleander picks the machine from how much a query reads. There is no per-query fee and no charge to start or stop.
- Spark and Bloom cost $0.0000175 per vCPU second and $0.0000019 per GB second. This estimate holds every machine for the whole run. Real runs bill measured seconds, so they can come in lower. Iceberg maintenance is a Spark job, so it bills at these rates too. Queries oleander routes to Spark still count toward your plan's maximum query input.
- Distributed Polars costs $0.00006 per vCPU second and $0.000007 per GB second across its workers and a 2 vCPU coordinator, from cluster startup until the run finishes.
- A month is 30 days. Hobby is never invoiced: queries stop once a month's compute reaches $5. Estimates exclude tax and anything you pay AWS for your own bucket. See all plans and rates.
Best fit
Pro
$372 / mo