All Tools View Categories About Contact Privacy

Cluster Capacity Calculator

Estimate nodes needed from total workload and node capacity.

Runs entirely in your browser - nothing is uploaded and no cluster connection is made.
Your plan will appear here.

About Cluster Capacity Calculator

The Cluster Capacity Calculator estimates how many worker nodes you need to schedule a total workload. Enter the total requested CPU (cores) and memory (GiB), the capacity of one node, and a headroom percentage.

It computes the nodes required for CPU and for memory separately and reports the binding resource that drives the total node count.

Nothing is uploaded. The calculation runs entirely in your browser.

Features

  • Per-resource math - CPU and memory nodes.
  • Headroom - reserve margin per node.
  • Binding resource - what drives the count.
  • Copy / Download / Print - export the plan.

How to Use

  1. Enter total CPU and memory requests.
  2. Enter one node capacity.
  3. Set headroom.
  4. Click Calculate.

Examples

Example 1 - no headroom. 4 cores / 16Gi on 2-core / 8Gi nodes needs 2 nodes.

Example 2 - 20% headroom. same load now needs 3 nodes.

Benefits

  • Right-size - avoid over/under provisioning.
  • Headroom aware - realistic packing.
  • Private - runs fully in the browser.
  • Exportable - copy, download or print.

Frequently Asked Questions

What does this calculator estimate?
Given the total CPU and memory your workloads request and the capacity of a single node, it estimates how many nodes are required, per resource and overall.
What is headroom?
A safety margin (for example 20%) reserved on each node for system daemons and burst, reducing the schedulable capacity used in the math.
Which resource drives the count?
The binding resource - whichever of CPU or memory needs more nodes becomes the overall node count.
Does this connect to a cluster?
No. Enter your numbers; everything is computed in your browser.
Are these exact?
It is a planning estimate assuming perfectly packable pods. Real packing depends on pod topology and limits.