Skip to content

Resource pricing

CruiseKube surfaces approximate cost and savings in the dashboard by applying hourly unit rates to CPU and memory. This is a heuristic, not billing integration.

Beta

Cost monitoring is beta. CruiseKube does not maintain a full time-series billing history; figures are point-in-time estimates from current recommendations and requests.


Mental model

In the real world you pay for instances (on-demand, spot, reserved, etc.). The dashboard instead uses:

  • A configurable $/CPU core / hour
  • A configurable $/GB memory / hour

All “monthly” numbers use 720 hours/month unless noted otherwise in the product.


Default assumptions (built-in reference model)

When no custom rates are set, calculations use a reference derived from public EC2 pricing:

  • Base shape: AWS c5a.xlarge (4 vCPU, 8 GiB).
  • Blend: half on-demand, half spot (no reserved).
  • Split: half of instance $/hr attributed to CPU, half to memory (network/storage bundled for this model).
  • GPU instances are excluded from this reference path.

From instances.vantage.sh — c5a.xlarge (illustrative; prices change):

  • Blended instance \(/hr ≈ **(\)0.154 + \(0.078) / 2** → **\)0.116/hr**
  • Per vCPU before split: \(0.116 / 4** → **\)0.029/core/hr
  • Per GiB before split: \(0.116 / 8** → **\)0.0145/GiB/hr
  • After 50/50 CPU vs memory attribution on that synthetic split, effective defaults used in-product are approximately:
Resource Default (USD)
CPU 0.0145 / core / hour
Memory 0.00724 / GiB / hour

Where to configure

  1. Open the dashboard (see Dashboard).
  2. In Policies & Configuration, open the Resource pricing (or equivalent) tab—CPU and memory unit rates used for dashboard cost cards and per-workload savings.
  3. Enter CPU and memory unit prices.

Values are stored in the browser for the cost views (they are not a substitute for your cloud bill).

Resource pricing settings in Policies & Configuration



Honest limitations

  • Does not know your committed use, enterprise discounts, or regional price tables.
  • Does not replace cloud cost tools (CUR, FinOps platforms).
  • Savings are “if recommendations were fully applied,” not a bankable invoice line.

Use the numbers for directional decisions and executive storytelling—then validate against your real CUR or billing export.