Pod Count Calculator
How many replicas of a service to run for a peak request rate — from what one pod handles, the utilization you want to run at, the minimum for resilience and the zones to spread across — with the CPU and memory that adds up to.
Divide the peak request rate by what one pod sustains at your target utilization (running at 60–70% leaves room for spikes and for a pod dying), round up, and never go below the minimum you need for resilience.
How the pod count calculator works
Divide the peak request rate by what one pod sustains at your target utilization (running at 60–70% leaves room for spikes and for a pod dying), round up, and never go below the minimum you need for resilience. Spread across availability zones so losing one zone does not drop capacity below the peak: that means enough pods that the remaining zones still cover it. Multiply by the per-pod requests for the cluster resources.
Formula: pods = max(min, ceil(peak ÷ (per-pod capacity × utilisation))); zone-safe pods = ceil(pods × zones ÷ (zones − 1))
Worked examples
| Inputs | Replicas to run | Note |
|---|---|---|
| 2,400 req/s, 150 per pod, 65%, three zones | 38 | 25 for the load, 38 to survive a zone |
| A small service under its minimum | 5 | 5 — the zone rule on a floor of 3 |
| Single zone, 1,000 req/s at 80 per pod | 18 | 18 |
FAQFrequently asked questions
Why not run pods at 100%?
Because load is spiky and pods die: at full utilization any bump queues requests and any failure overloads the rest. 60–70% leaves the autoscaler time to react and the survivors room to absorb a lost pod.
Why the zone arithmetic?
If a third of your pods are in a zone that goes down, the other two thirds must still handle the peak. Sizing so that (zones − 1) ÷ zones of the pods cover the load is the usual rule; with two zones it doubles the count.
How do I find the per-pod capacity?
Load-test one pod to the point where latency climbs, and take 100% as the rate just before that. It depends on the CPU request, so re-measure if you change it.
What about the autoscaler?
Use this figure as the minimum replicas for the busy period, or as a sanity check on the HPA’s target utilization and maximum. Autoscaling reacts after the fact; capacity planning is for the load you know is coming.
Where these figures come from
- IEC 80000-13 — Information science and technology (quantities and units) — the decimal (kB, MB) versus binary (KiB, MiB) prefixes used throughout
- RFC 4632 — Classless Inter-domain Routing (CIDR) — the address-plan arithmetic behind the subnet calculator
- RFC 1918 — Address Allocation for Private Internets — the private ranges the subnet calculator recognises
- NIST SP 800-63B — Digital Identity Guidelines, Authentication — length over composition rules; the basis of the password guidance here
- NIST SP 800-57 Part 1 — Recommendation for Key Management — key-strength comparisons used by the key-space calculator
- NIST — Digital Identity Guidelines (SP 800-63) — US federal authentication standard
Last checked: September 2026. Units follow the SI decimal convention (IEC 80000-13 defines the binary alternatives); network and security figures cite the defining standard.