Illustrative one-region-loss capacity requirement at 18,000 requests/s
339 GPUs in three regions; 450 in twoderived
| Value kind | derived — Derived values are guide calculations; their result is only as fixed as the stated inputs, scope, and method. |
|---|---|
| Scope | Assumed 100 requests/s/GPU and 80% loading limit; equal regions, compatible GPUs, one-region-loss protection. r*ceil(18000/((r−1)*100*0.8)); capacity ratios are not cost ratios. |
| As of | 2026-09 |
| Source | Guide derivation from exact illustrative inputs; AWS supplies DR pattern boundaries only, not GPU throughput or counts. · Worked regional capacity |
| Review | checking…review by 2027-09-05 · standard cadence |
| Recorded changes | last 2026-09-16 |
| Claim id | part12-dr-capacity-case |
Where the guide uses it
← Full numbers register — every date-stamped figure in the guide, with revision history.