Minimum resource requirements¶
Use these values as the minimum baseline when you plan AI Gateway capacity.
Minimum baseline for Kubernetes¶
The production deployment guide provides these baseline requests and limits.
| Component | CPU request | Memory request | CPU limit | Memory limit |
|---|---|---|---|---|
| Gateway Controller | 500m |
1Gi |
1000m |
2Gi |
| Gateway Runtime | 2000m |
2Gi |
4000m |
2Gi |
For high availability, keep two replicas of each component.